CA2232534C - Method and apparatus for determining the alignment of motor vehicle wheels - Google Patents

Abstract

An apparatus for determining the position and alignment of wheels, including targets (130) for attachment to the wheels (112-115), a pair of optical sensing means such as television cameras (122, 124) for viewing the targets, an electronic processing means connected to the optical sensing means for processing data relating to images of the targets to determine position and alignment information, and a display means for displaying the position and alignment information. The optical sensing means view a target and form an image. Electronic signals corresponding to each of the images are transferre d to the electronic processing means which correlates the image signals of eac h of the targets with the true shape of each target. The processing means relates the geometric characteristics and positional interrelationships of certain known elements of the target with corresponding elements in the view ed images and calculates the position and alignment of the wheels to which the targets are attached.

Description

SPECIFICATION
"Method and Apparatus for Determining the Alignment of Motor Vehicle Wheels°
BACKf3R0UND OF T8E INVENTION
Field of the Inveation This invention relates to a method and apparatus for determining the alignment of motor vehicle wheels. More particularly, this invention relates to a method and apparatus including an opto-electronic image detector for detecting wheel orientation and producing electronic image data representing the wheels, or a reference mounted thereon, computational means for determining the alignment of the wheels, and means for comparing the electronic images, or data corresponding thereto, to previously stored alignment data and generating information which can be used to perform necessary adjustment to the vehicle.
Terms and Definitions In the vehicle wheel alignment industry the following terms, with corresponding definitions, are commonly used:

CA 02232534 1998-03-19 "
4 ' . w0 97/I4016 PCT/LTS96/16362 _ -2-1 Camber is the angle representing the inward or .

2 outward tilt from true vertical of the wheel and is 3 positive if the top of the wheel tilts outward.

4 . Caster is the angle representing the forward ox , rearward tilt from true vertical of the steering axis.
6 When a wheel is viewed from the side, the angle is _ 7 positive when the upper ball joint (or top of king pin, 8 or upper mount of a McPherson strut) is rearward of the 9 lower ball joint (or bottom of the king pin, or lower mount of a McPherson strut).
11 Thrust Line (T/L) is a line that bisects.the angle 12 formed by the rear toe lines. Lines and angles measured 13. clockwise from the 12:00 axis are positive.
14 Geometric Center Line, is the line that runs from a 15~ point on the rear axle midway between the rear wheels to 16 a point on the front axle midway between the front 17 wheels.
18 Individual Toe is the angle formed by a front-to-19 back line through the wheel compared to the geometric center line. Angles pertaining to the left side are 21 positive when clockwise of the thrust line and angles 22 pertaining to the right side are positive when 23 counterclockwise of the thrust line.
24 Offset is the amount that a front wheel and its 25_ corresponding rear wheel are out of line with each 26_ other. If there is no offset, the rear wheel is 27 directly behind the front wheel.
28 Setback is the amount that one wheel on one side of 29 the vehicle is displaced back from its corresponding wheel on the other side of the vehicle.
31 Steering Axis is a line projected from the upper 32 pivot point of the upper ball joint or top of kingpin, 33 or McPherson strut, through the lower ball joint. , 34 Steering Axis Inclination (SAI) is the angle between the steering axis and true vertical. If the 36 steering_axis appears to tilt inward at the bottom of 37 the wheel (as viewed from the driver position), the SAI
38 is positive. SAI also is also_known as kingpin 39 inclination (KPI).

WO 97/14016 , PCT/US96/I6362 ,.1., . , Thrust Angle (T/A) is the angle between the thrust 2 line and the geometric center line. Angles measured 3 clockwise from the.geometric center line are positive. , 4.,~....,_ Total Toe is the sum of individual, side-by-side 5, toe, measurements. If lines projected parallel to the 6. primary planes of the wheels intersect at a point ahead 7 of the side-by-side wheels, the angle is positive (toe 8 in). If the lines would intersect behind the side to 9 side wheels, the angle is negative (toe out). If the projected lines are parallel, the toe is zero.
11 Traditionally, the Camber and Toe measurements for 12 each wheel of the vehicle are relative measurements i.e.
13 relative to a vertical plane or to another wheel and 14 these measurements are therefore made when the wheels are stationary. On the other hand, the calculation of 16 Caster and SAI is a dynamic procedure and entails 17 determining how the Camber of the front wheels changes 18 with respect to a change in steering angle. This is 19 usually done by swinging the front wheels from left to right through an angle of between 10° and 30°, or vice 21 versa, while determining the resultant changes in Camber 22 of the wheel with steering angle changes. From these 23 determinations the Caster and SAI are determined by 24 methods well known in the wheel alignment industry.
Similarly, once Camber, Toe, Caster and SAI have 26 been measured, all other relevant wheel alignment 27 parameters can be calculated by methods and formulations 28 well known in the industry.

Brief Description of the Prior Art 31 The wheels of a motor vehicle need to be 32 periodically checked to determine whether or not they 33 are in alignment with each other because, if any of the 34 wheels are out of alignment, this can result in -, 35 excessive or uneven wear of the tires of the vehicle 36 and/or adversely affect the handling and stability of 37 the vehicle.
38 The typical steps of determining and correcting the 39 alignment of a vehicle's wheels are as follows:

CA 02232534 1998-03-19 -% , ' W0 ~97/140I6 PCT/US96/16362 ,1 1. The vehicle is driven onto a test bed or rack 2 which has previously been levelled to ensure a level 3 _. base for the vehicle.
4_ _ 2. Some components of the alignment determination _ y apparatus are mounted onto the wheels of the vehicle.
6 These components are not necessarily accurately placed 7with respect to the wheel axis. The extent of the 8 inaccuracy by which these components are mounted is 9 called the "mounting error".
3. A "runout" calculation is done by jacking the 11 vehicle up and rotating each wheel and taking 12 measurements of the orientation of that wheel at 13 different positions. These measurements are then used 14. to calculate a correction factor to compensate for the "mounting error" and actual rim run-out.
16 4. A determination of the alignment of each of 17 the wheels is done. The results of these determinations 18 are compared to the specifications of alignment 19 parameters for the vehicle being tested.
5. The operator then adjusts the various linkages 21 of each wheel to correct for the misalignment, if any, 22 of the wheels.
23 6. Steps 4 and 5 are repeated until the alignment 24 is up to standard and/or is within manufacturer's specifications.
26 A large variety of devices for measuring the 27 alignment of a motor vehicle's wheels exist. Many of 28 these use optical instrumentation and/or light beams to 29 determine the alignment of the wheels. Examples can be found in United States Patent Nos. 3,951,551 31 (Macpherson); 4,150,897 (Roberts); 4,154,531 (Roberts);
32 4,249,824 (Weiderrich); 4,302,104 (Hunter) ; 4,311,386 33 (Coetsier); 4,338,027 (Eck); 4,349,965 (Alsina);
34 4,803,785 (Reilly) and 5,048,954 (Madey).
All these devices operate with an apparatus which 36 is mounted onto the wheel of a vehicle and which emits ' 37 or reflects a light beam to illuminate an area on some 38 form of reference such as a reference grid. As the 39 position of the area illuminated by the beam on the t CA 02232534 1998-03-19 -S-1 reference is a function of the deflection of the beam, 2 ,.,which in turn is a function of the orientation of the ' 3 wheel, the alignment of the wheel can be calculated from , 4:. the positioning of the illuminated area on the reference.
6 Other devices utilize a measuring head mounted onto each wheel of the vehicle. These heads typically 8 include gravity gauges that are either connected to 9 adjacent heads by means of cords or wires under tension or, alternatively, configured with beams of light 11 shining between adjacent heads. The measuring heads, 12 which must be maintained level, are then able to measure 13 the relative angles between adjacent cords/beams of 14 light as well as the angles between each wheel and its adjacent cord/beam of light and, from these 16 measurements, calculate the alignment of the wheels.
17 Another type of alignment device is illustrated in 18 United States Patents 4,899,218 (Waldecker) and 19 4,745,469 (Waldecker et al). This device operates by projecting structured light onto a wheel of the motor 21 vehicle so that at least two contour lines on the 22 surface of the wheel are illuminated. These contour 23 lines are then read by video cameras which are 24 positioned offset from the optical plane of the structured light and which are connected to a processor 26 which calculates the spatial position of the contour 27 lines (and therefore that of the wheel) by means of 28 triangulation.
29 Generally, the heads used in the above described wheel alignment devices are delicate and expensive, 31 complicated to use and must be carefully set up.
32 Furthermore, certain of these devices rely on the . 33 accurate placing of optical or other measuring devices 34 either on or in a set position relative to the wheels of . 35 the vehicle. This can be time consuming and complicated 36 for the technicians operating the alignment 37 determination apparatus. Such equipment also has the 38 disadvantage that components which are carelessly left 39 secured to the wheels when the vehicle is moved from the CA 02232534 1999-11-17 .
WO 97/14016 PCT/US96/16362 .

l.. test area can very easily be damaged. Such damage, 2 . particularly in the case of sophisticated equipment, can 3 be costly.
4 German patent application DE 29 48 573*in the name _ of Siemens Aktiengesellschaft discloses an apparatus 6 which can be used to determine both the orientation and _ the spatial position of the plane of the wheel of a 8 motor vehicle as well as the three-dimensional position 9 of the steering axis of.this wheel. The application discloses a method whereby a television camera takes an 11 image of the rim on the wheel from two different known 12 height positions. These images are fed into a processor 13 which relates them to the known coordinates and viewing 14 angles of the camera at its two height positions and determines the three-dimensional position of the rim.
16 In a similar way, a number of images of each wheel, 17 in different steering positions, are taken to determine 18 a three-dimensional solid of revolution for the wheel.
19 From the axis of this solid of revolution the steering axis of the wheel under investigation can be determined.
21 As a result, the three-dimensional position of both the 22 steering axis and the center point of the plane defined 23 by the rim of the wheel is determined.
24 In addition to the fact that little indication is given as to how the above values are determined, the 26 method and apparatus of the described application has 27 the disadvantage that, because a triangulation technique 28 is used, at least two images (from different cameras or 29 from a single camera viewing along different axes) of the wheel must be taken. Furthermore, both the 31 coordinated three-dimensional position for each point 32 from where an image of the wheel is taken as well as the 33 orientation of each of the view paths must be accurately 34 known.
This is a major disadvantage of this invention 36 because the accurate determination of the three 37 dimensional positions and the orientation of the view 38 paths, requires sophisticated equipment which can easily * published June 1981 _7_ go out of calibration due to temperature changes, vibration, ground movement, etc.
A further disadvantage is that the method in this applica-tion does not indicate how it makes allowances for the perspec-tive distortion of the image of the rim of the wheel. This perspective distortion causes the image of the rim to be in the form of a distorted ellipse with the edge of the ellipse closest to the television camera appearing larger and the image of the edge farthest from the camera appearing smaller. If allowance for this distortion is not made, inaccuracies can result.
The need therefore still exists for a wheel alignment apparatus which is simple and easy to use, which has its sophisticated alignment detection components remote from the wheels of the motor vehicle and which can provide reliably accurate alignment measurements over a large range of rim diameters, track widths and wheel bases.
SD1~1ARY OF TH8 INVENTION
Accordingly the invention seeks to provide a wheel alignment apparatus which is simple, easy and quick to use.
Further the invention seeks to provide a wheel alignment apparatus which can operate with its precision components removed from the motor vehicle.
Still further the invention seeks to provide a wheel alignment apparatus which uses an opto-electronic image detection device to determine the alignment the wheel.
Further still the invention seeks to provide a wheel alignment apparatus which uses a perspective image, of a known target attached to a wheel, to determine the orientation of the target and thereby the alignment of the wheel.
The invention in one broad aspect provides an apparatus for determining the alignment of motor vehicle wheels comprising target means including at least first and second target objects for attachment to wheels on respective first and second sides of a vehicle under inspection, each target object including a plurality of visually perceptible, geometrically configured target elements having known geometric characteristics and positional interrelationships. Optical inspection means defines _8_ a spatial reference system and includes first and second cameras respectively forming first and second viewing paths intersecting the first and second target objects when they are attached to wheels of a vehicle under inspection, each camera being operable to inspect an image of a corresponding target object as viewed along its corresponding viewing path and to generate image information describing the geometric characteristics and positional interrelationships of the imaged target elements.
The apparatus has processing means for relating such image information to predetermined reference information describing the known geometric characteristics and positional interrelationships of the target elements and for determining the position and angular orientation of the first and second target objects relative to the spatial reference system and for generating first and second position and orientation information commensurate therewith. Means responsive to the first and second position and orientation information is operative to indicate the position arid alignment of the wheels to which the first and second targets are attached.
The invention also pertains to a method of determining the relative orientation of the wheels of a vehicle comprising the steps of establishing a first target means on a first wheel on a first side of the vehicle and establishing a second target means on a second wheel on a second side of the vehicle, each target means including a plurality of target elements of known geometric characteristics and positional interrelationships, the first and second target means having a predetermined positional relationship to the wheels to which they are attached. A first camera means having a known spatial position and directional orientation is used to view the first target and to form a first detected image thereof and a second camera means having a known spatial position and directional orientation is used to view the second target to form a second detected image thereof. The geometric characteristics and positional interrelationships of the target elements of each of the first and second detected images is determined and the determined geometric characteristics and positional interrelationships of the first detected image is related to the known geometric characteristics and positional interrelationships of corresponding elements of the first target means to determine the angular orientation of the first target means. The determined geometric characteristics and positional interrelationships of the second detected image is related to the known geometric characteristics and positional interrelationships of corresponding elements of the second target means to determine the angular orientation of the second target means and the angular orientations are used to determine the alignment of the first and second wheels.
More particularly a presently preferred embodiment of this invention includes an apparatus for determining the aligrunent of a motor vehicle's wheels and comprises an optical sensing means such as a television camera, an electronic processing means connected to the optical sensing means, at least one predetermined target which either forms part of the wheel or is attached thereto and a display for indicating the detected alignment. The optical sensing means views a target attached to each wheel and forms a perspective image of each target.
Electronic signals corresponding to each of the images are transferred to the electronic processing means which correlates the perspective image of each of the targets with the true shape of each target. In so doing, the processor relates the dimensions of certain known geometric elements of the target with the dimensions of corresponding elements in the perspective image and by performing certain trigonometric calculations (or by any other suitable mathematical or numerical methods), calculates the alignment of the wheels of the vehicle. This invention can also be used to calculate the three-dimensional position and orientation of the axis of rotation of the wheel (wheel axis). The detected alignment is then displayed for use in performing alignment adjustments to the vehicle.
Preferably, the optical sensor means forms images of a target attached to each of at least two wheels mounted on the same axle of the vehicle and the electronic processing means calculates the relative angles between the two wheels. Even more preferably, the optical sensor means forms images of all the targets on the wheels and relative alignment calculations are computed by the electronic processor means for all these images.

CA 02232534 1998-03-19 .
. WO 97/14016 PCT/US96/16362 _g_ 2 - ... FIGS. 1 (a) - (c) . illustrate three different images of 3 a circle resulting.from various degrees of rotation , 4 about different axes;
~~ ,_-_FIG. 2,is a schematic representation illustrating ~6 ,the-apparatus and method of the invention;
FIG. 2a is an illustration of a quasi three-8 dimensional representation of a type that may be 9 generated on a system display screen to report detected alignment and to guide the technician in making 11 appropriate vehicle adjustments;
12 FIG. 2b is a cross-section through a pan-and-tilt 13 mirror used in one embodiment of this invention;
14 FIG. 3 is a representation of an exemplary target that can be used with the apparatus in FIG. 2;
16 FIG..4 is a schematic representation of an 17 alternative embodiment of the apparatus of this 18 invention;
19 FIG. 5 is a perspective view of an alternative target mounted on a vehicle's wheel;
21 FIG. 6 is a schematic representation of an image of 22 the target illustrated in FIG. 5 formed by using the 23 optical system in FIG. 4;
.24 FIG. 7.illustrates one method of how the apparatus calculates the run-out factor of the wheel;
26 FIGS..8a-8c illustrate certain aspects of the 27 mathematics performed in the method and apparatus of 28 this invention;
29 FIG. 9 is a diagram schematically illustrating another alternative embodiment of the present invention;
31 FIG. 10 illustrates details of the camera/light 32 subsystem of FIG. 9; and 33 FIG. 11 illustrates an alternative embodiment of a 34 target array..
36 DESCRIPTION OF. PREFERRED EMBODIMENTS
37 Basic Theory of the Invention 38 This invention is based on the fact that the 39 image of a body varies according to the perspective from a , 1 which such body is viewed and that the variation in the ,.2 image is directly related to and determinable from the 3 perspective angle of the view path along which the body 4 is viewed. -., ,_, Furthermore it is known that it is possible to 6 determine the perspective angles at which an object is '7 viewed merely by relating the perspective image of that 8 object with a true non-perspective image thereof.
9 Conversely put, it is possible to determine the angles at which an object is orientated to a view path (or a 11 plane perpendicular thereto) by comparing a perspective 12 image of an object with a non-perspective image thereof.

13 .This is illustrated in FIGS. 1(a)-(c) with 14.. reference to a circle 10, shown as it would appear if viewed from three different perspectives. In FIG. 1 (a) 16 the circle 10 is illustrated as it would appear if it 17 were viewed along an axis perpendicular to its primary 18 plane which, in this case, is in the plane of the paper.
19 If--this-circle-is rotated through an angle 6, being less than 90°, about the y-axis 12 and viewed along the same 21 view path, the image of the circle 10 will be that of an 22 ellipse as shown in FIG. 1(b). Similarly, if the circle 23 is rotated about both the x and the y-axes, 12 and 14 24 respectively, through angles 9 and ~ respectively, the image of the circle (the ellipse) will be as shown in 26 FIG. 1(c), in which the major axis 16 of the ellipse is 27 shown to be angled relative to both the x and y-axes.
28 It will, however, be realized that the ellipses 29 here are idealized in that they make no allowance for the distortion which results in an image when it is 31 viewed from a perspective angle. This distortion is 32 illustrated by the broken lines 11 in FIGS. 1(b) and 33 (c). As can be seen from these Figures, the edge of the .
34 ellipse 11, which is closer to the viewer, appears larger while the edge 11, which is farther from the 36 viewer, appears smaller. The resulting image 11 is thus 37 a distorted ellipse.
38 Returning to the idealized conditions shown in 39 these figures, and assuming the angles A and ~ are not WO 9!7/14016 PCT/I1S96/16362 1._ known, it is possible to determine the orientation of 2 the primary plane of the ellipse illustrated in FIG.
. 3.:,.'- 1 (c) by relating the image of the ellipse to the circle 4 10 in FIG. 1(a). This is usually done by.relating the ~ geometric characteristics (e. g. dimension) of at least 6 one element of the ellipse (e. g. the major and minor 7 , axes 16, 18 thereof) to characteristics of corresponding 8 elements (the diameters) of the circle in FIG. 1(a).
9 Under idealized conditions, these orientation calculations~are done by applying trigonometric 11 functions or any other mathematical/numerical methods to 12 the ratios between the minor and/or major axis and the 13 diameter. In addition, the angles of the minor and 14 major axes to the horizontal (x-) axis or vertical (y-) axis can be calculated. Once all these angles have been 16 determined, the orientation in space of the primary 17 plane of the ellipse will be determined.
18 Although not illustrated, it is also possible to 19 determine the position in space of the circle 10. This will, however, be demonstrated below with reference to 21 FIG. 8.
22 The performance of the above illustrated 23 calculations is complicated by the real-life perspective 24 distortion of the image, as illustrated by the broken lines 11. How this foreshortening is allowed for will, 26 once again, be discussed with reference to the 27 mathematics illustrated in FIG. 8.

29 Brief Description of One Embodiment of the Alignment Apparatus of the Invention 31 The apparatus with which this theory is applied in 32 this invention is illustrated in the schematic 33 representation in FIG. 2. In this figure a motor 34 vehicle 20, on which a wheel alignment is to be performed, is represented by a schematic illustration of 36 its chassis and is shown to include two front wheels 22L
37 and 22R and two rear wheels 24L and 24R. The vehicle 20 38 is shown positioned on a conventional wheel alignment ~.~, _ WO X7/14016 PCT/US96/I6362 a , 1 " test bed 26, indicated in dotted lines, which does not 2 form part of this invention.
3- . The alignment apparatus of this invention is shown 4 to be constituted by a video camera 30 which is in 'electrical communication with an electronic processing .6~ means such as a computer 32 which, in operation, 7displays results and calculations on a visual display 8 unit 34. In addition, the apparatus includes a keyboard 9 36 (or some other suitable means) for inputting data and relevant information into the computer 32. It will, of 11 course, be appreciated that display and keyboard entry 12. could be provided by a remote unit which communicates 13 with the computer through a cable, lightwave or radio 14 link.
In accordance with a preferred embodiment and as 16 illustrated in FIG. 2a, a computer-generated quasi 17 three-dimensional representation of the wheels being 18 aligned may be depicted on the display unit 34 along 19 with suitable indicia evidencing the detected alignment.
In addition, alphanumeric and/or pictorial hints or 21 suggestions may be depicted to guide the technician in 22 adjusting the various vehicle parameters as required to 23 bring the alignment into conformance with predetermined 24 specifications.
The video camera 30 sights onto the wheels 22L, 26 22R, 24L and 24R along a view path 38 which passes 27 through a lens 40 and onto a beam splitter 42. The beam 28, splitter 42 splits the view path 38 into two components 29 38L and 38R, respectively. As is apparent from this figure, the left hand component 38L of the view path 38 31 is reflected perpendicularly to the initial view path by 32 the beam splitter 42 while the right hand component 38R
33_ is reflected perpendicularly to the initial view path by 34 a mirror or prism 44 mounted adjacent to the beam splitter. The apparatus also includes a housing 48 into 36 which the beam splitter 42, mirror 44 and at least two 37 pan-and-tilt mirrors, 46L and 46R, are mounted. From 38 this point onwards the respective components of the 39 apparatus and the view path are identical for both the WO 97/14016 , PCT/US96/16362 .i ".left and right,side of the motor vehicle and therefore a 2 .description of. only one side will suffice.
3 The.left hand-component. of the view path 38L is 4 , reflected. onto ,the wheels. 22L and 24L. by the left side 5. 'pan-and-tilt mirror.46L which is movable to allow the 6 video camera 30 to consecutively view the front wheel 7 22L and the rear wheel 24L of the vehicle 20. In some 8 embodiments of this invention the pan-and-tilt mirror 9 46L can be configured so that both the front and rear wheels of the motor vehicle can be viewed 11 simultaneously.
12 In this embodiment, the view path 38L passes from 13 the pan-and-tilt mirror 46L through an aperture 50L in 14 the wall of the housing 48 and onto the respective wheels 22L and 24L. A shutter 52L is positioned so that 16 it can be operated to close the aperture 50L thereby 17 effectively blocking the view path 38L and allowing the 18 video camera 30 to sight onto the right hand side of the 19 vehicle 20 only. Alternatively, shutters could be placed at the locations 53L and 53R and/or an electronic 21 shutter within the camera 30 could be synchronized with 22 one or more strobed light sources to permit capture of 23 an image only when a particular target or targets are 24 illuminated.
26 Operation of the Alignment Apparatus 27 In a typical operation, the apparatus of this 28 embodiment of the invention works as follows: The 29 vehicle 20 is driven onto the test bed 26 which basically consists of two parallel metal strips on which 31 the wheels of the vehicle rest. Under the test bed, a 32 lift mechanism is located (but not shown) which acts to 33 lift the metal strips and the vehicle to allow the wheel 34 alignment technician to access the wheel mountings to correct misalignment of the wheels. In addition, a 36 rotationally mounted circular plate commonly called a 37 turnplate (not shown), is located under each front wheel 38 of the vehicle. The turnplates allow the front wheels 39 to be pivoted about their steering axes relatively ~~,F. WO 97/14016 PCTlCTS96/16362 1 easily. This facilitates the procedure involved during 2 the calculation of caster and other angles determined 3._ dynamically. The rear wheels are positioned on .
<"4,__'elongate, rectangular, smooth metal plates mounted on _ the metal strips. These plates are usually termed skid 6 plates and allow the rear wheels to be adjusted by a 7 technician once the rear wheel mountings have been 8 loosened. Such plates also prevent preload to wheels 9 tending to affect their~angular position.
In addition, as in some sophisticated alignment 11 machines, the vehicle make and model year can be entered 12 into the apparatus at some time early on in the 13 procedure, and this information is used by the apparatus 14 to determine the alignment parameters, for the vehicle concerned, from previously programmed lookup tables 16 within the computer 32. Furthermore, from the vehicle's 17 make and model year, the track width and wheelbase 18 dimensions can be determined by retrieving the data from 19 memory. These can be used to drive the mirrors of the alignment apparatus to "home" in on the wheels of the 21 vehicle more accurately. Alternatively, previous 22 operational history information can be used to select 23 likely wheel location. Still another possibility is to 24 cause the mirrors to sweep a particular pattern.
Once the vehicle 20 has been driven onto the test 26 bed 26, a target 54 is mounted onto each wheel. The 27 shape and configuration of the target will be described 28 later with reference to FIG. 3. The apparatus first 29 makes a "run-out" factor calculation according to the method that will more fully be described with reference 31 to FIG. 7.
32 Once the "run-out" factor has been calculated, the 33 alignment apparatus forms an image (a detected image) of 34 each of the targets 54 on the wheels of the motor vehicle 20. These detected images are.processed in the 36 electronic processing means/computer 32 which 37 calculates, using the method of the invention as will be 38 more fully described, the orientation of each of the 39 targets to the respective view paths 38L, 38R. The ~ CA 02232534 1998-03-19 WO 9?/14016 PCT/US96/16362 1 computer 32 then takes into account the "run-out"
2 factors mentioned above to calculate the true 3 ., orientation of the wheels relative to the respective .
4 view paths. Thereafter the apparatus makes allowance 5__. for the orientation of the pan-and-tilt mirrors 46L, 46R
6 to calculate the actual orientation of the primary 7 . planes of each of the wheels. Upon this being done, the 8 results of the computation are displayed on the display 9 34 which gives the operator the required instructions as to which corrections need to be made to, for example, 11 adjustments to the steerage linkages 60 of the front 12 wheels 22L and 22R to correct the detected misalignment 13 of the wheels of the vehicle.

Orientation Calculations 16 The computer 32 does all the required calculations 17 using a computer program such as IMAGE ANALYST, which is 18 capable of analyzing images and values associated 19 therewith. Typically, IMAGE ANALYST produces values for the..center points of these images in coordinates 21 relating to the pixels on the screen of the video 22 camera. These values are then processed by software 23. which incorporates the later-to-be-described mathematics 24 illustrated with respect to FIG. 8. Although software such as IMAGE ANALYST may have many features, in this 26 application it is apparent that the main features 27 utilized in this application is that of being able to 28 provide screen coordinates for the images detected by 29 the video camera. It is, therefore, possible for software other than IMAGE ANALYST to be used with this 31 method and apparatus. IMAGE ANALYST is supplied by 32 AUTOMATIX, INC. of 755 Middlesex Turnpike, Billerca, MA
33 01821.

Orientation of the Pan-and-Tilt Mirrors 36 In the above-described method it is evident that 37 knowledge of the orientation of the pan-and-tilt mirrors 38 46L, 46R is required for the effective calculation of 39 the relative alignment of the wheels of the vehicle 20 . WO 97/14016 PCT/US96116362 v 1 to. each other. The orientation of these mirrors 46L, 2 46R can be determined iri one of two ways. One way of 3:determining the orientation is by linking the mirrors 4~__46L, 46R to a sensitive tracking and orientation -5.~~~~determination device which outputs data to the computer .. r =L~:~..
6 32 which, in turn, calculates the orientation of the 7 mirrors in.three-dimensional space. Alternatively, and 8 preferably, the face of each mirror includes a clearly 9 defined pattern, usually in the form of a number of . small, spaced-apart dots, which define an identifiable 11 pattern that can be detected by the video camera 30 as 12, it sights onto the wheels of the motor vehicle 20. Once 13 the video camera 30 has detected the pattern on the 14 mirrors 46L, 46R it can form an image thereof; an image which, because of the orientation of the mirrors, will 16 be a perspective image, and which can then be 17 electronically fed into the computer which, in turn, can 18 calculate the minor orientation in three-dimensional 19 space along the same lines as the orientation of the wheels of the vehicle 20 are calculated. This second 21 alternative is preferable because it does not require 22 sophisticated and expensive electronic tracking and 23 orientation determination equipment.
24 One way of implementing this second, preferable alternative, is to incorporate a lens 40 into the 26 apparatus. The lens has a focal length such that it 27 projects an adequately clear image of both the targets 28 and the mirrors onto the camera 30.
29 In FIG. 2b, one way of enhancing the images of the dots on the pan-and-tilt mirrors is illustrated. This 31 figure illustrates a cross-section through a pan-and-32 tilt mirror 46L with two dots 41 shown formed on its 33 upper surface. A plano-convex lens 43 is located on top 34 of each dot. The focal length of each of these lenses is such that, together with the lens 40, they form a 36 clear image of the dots in the video camera 30.
37 Although this figure illustrates two individual plano-38 convex lenses 43, it will be evident that a single lens . ~ CA 02232534 1998-03-19 WO 97/14016 ' PCT/US96/16362 _1,7_ 1 spanning two or more dots could be used. Similarly, 2 other optical methods can be used to accomplish this.
3 _ _ ..
4 Oriexztation of the Tarctets An example of a typical target 54 that can be used 6 on the wheels of the vehicle 20 is illustrated in FIG.
7 3. As can be seen from this figure, the target consists 8 of a flat plate with a pattern of two differently sized 9 circles 62, 63 marked in a pre-determined format thereon. Although a specific pattern is shown in this 11 figure, it will be evident that a large number of 12 different patterns can be used on the target 54. For 13 example, the target need not be circular, a larger or 14 smaller number of dots may be included. Moreover, other sizes and shapes can be used for the dots. In addition, 16 multifaceted plates or objects can also be used for the 17 targets.
18 In practice, a mathematical representation, or data 19 corresponding to a true image (i.e. an image taken by viewing the target perpendicularly to its primary plane) 21 and the dimensions of the target are preprogrammed into 22 the memory of the computer 32 so that, during the 23 alignment process, the computer has a reference image to 24 which the viewed perspective images of the targets can be compared.
26 The way that the computer calculates the 27 orientation of the target 54 is to identify certain 28 geometric characteristics on the target 54, take 29 perspective measurements of these and compare these measurements with the true image previously pre-31 programmed into the memory of the computer.
32 The apparatus could, for example, calculate the 33 center of each of the circles 62a,~ 62b by means of, say, 34 a method called centroiding. This is a method commonly used by image analysis computers to determine the 36 positioning of the center point or center line of an 37 object. Once the center points of the two circles 62a, 38 62b have been determined, the distance between the two 39 can be measured. This process is then repeated for CA 02232534 1998-03-19 ' WO l7/140I6 PCT/LTS96/I6362 1 other circles in the pattern on the target 54. These 2 distances can then be compared to the true distances 3 ~~ (i.e. non-perspective distances) between the respective 4 centers. Similarly, the angle to the horizontal (or -vertical) of the line joining the two centers can be J6-determined. Once allowance has been made for the effect _ 7 of the focal length of the lens 40 and other optical 8 characteristics of the components, such as beam splitter 9 42, mirror 44 and mirrors 46L, 46R, are considered, a calculation can be made as to what the orientation of 11 the target 54 is. This calculation can be done by using 12 trigonometric functions or other suitable mathematical 13 or numerical methods. As explained above, this will 14 also yield the orientation of the primary plane of the wheel of the vehicle.
16 . Although the above describes one method of 17 calculating the orientation of the target 54, it will be 18 evident that other methods are also available. For 19 example, the apparatus could sight onto only one of the circles, say the circle 63, and by using the perspective 21 image thereof (the distorted ellipse) calculate, in very 22 much the same way as described with reference to FIG. 1, 23 the orientation of that circle and, therefore, the 24 orientation of the target 54. Another example is to take two images rotated about 60° relative to each other 26~ and use such information to calculate the orientation of 27 the target with respect to its axis of rotation. Note 28 that only two images are required so long as the wheel 29 axle does not change its axial orientation. In addition, it is envisaged that in sophisticated 31 alignment systems more than one calculation will be 32 completed for each target and that the different results 33 of these calculations will be compared to each other to 34 ensure the required accuracy.
f' 1. Furthermore, as the true dimensions of the target 36 are preprogrammed into the memory of the computer 32, 37 the method and apparatus of this invention can be used 38 to determine the exact position of the wheels in three-39 dimensional space. This can be done by firstly WO 97!14016 PCT/US96/16362 1 determining the perspective image of certain of the 2 elements of the pattern on the target (for example, the 3 distances between circles) and comparing the dimensions , 4 of this image to the true dimensions of those elements.
This will yield the distance that the element and, 6 accordingly, the target 54 is from the video camera.
7 As the processes described above have already 8 yielded the orientation of target 54 with respect to the 9 view path and/or some other reference plane, this result can be combined with the calculated distance and the 11 geometric coordinates of the alignment apparatus to 12 yield a position of the target 54 relative to the 13 alignment apparatus. During this comparison process, 14 the effect of the focal length of the lens 40, as well as the optical characteristics of the beam splitter 42, 16 mirror 44 and the pan-and-tilt mirrors 46L and 46R must 17 also be taken into consideration. Typically, these 18 characteristics would be input into the computer by 19 direct e~:try or, preferably; by calibration techniques.
In this way the exact positioning of each~of the wheels 21 of the vehicle 20 can be calculated.

23 A Br~.ef Description of an Alternative Embodiment of the 24 Apparatus ofthe Invention It will be evident to one skilled in the art that a 26 number of different configurations of lens, beam 27 splitter and mirrors (i.e. the optical system) are 28 possible to achieve the required result with the method 29 and apparatus of this invention. One such configuration is illustrated in FIG. 4 of the accompanying drawings.
31 In this figure the equipment is shown to be 32 suspended over the motor vehicle 20 and includes a video 33 camera 30, computer 32 with associated display 34 and 34 data entry keyboard 36 as well as lens 40 similar to those illustrated in FIG. 2. As with the configuration 36 in FIG. 2, the view path or optical center line of the 37 video camera 30 is deflected into two directions 38L and 38 38R by a combination of beam splitter 42 and plane 39 mirror 44.

1 This configuration also includes two pan-and-tilt 2 mirrors 70L, 72L, located on the left side of the 3 apparatus and two pan-and-tilt mirrors 70R and 72R
4 located on the right side of the apparatus. The mirrors 70L, 72L are arranged to view the left front and left 6 rear wheels 22L, 24L, respectively and the mirrors 70R, 7 72R are arranged to view the right wheels 22R, 24R
8 respectively. As the mirrors 70L, 72L, 70R, 72R are 9 pan-and-tilt mirrors, they can be moved to view the wheels on the vehicle 20 even though the vehicle is not 11 accurately centered below the apparatus. These mirrors 12 are also useful in making allowance for vehicles of 13 different lengths of wheelbase and track width.
14 A further modification of this apparatus would include the replacement of the beam splitter 42 and the 16 plane mirror 44 with a single reflecting prism. The 17 prism has the advantage over the beam splitter 18 combination in that more light is reflected from the 19 prism into the camera 30. This results in a brighter image of the target 54 being formed by the camera 30.

22 Tarctet and.Taraet Image Details 23 With the apparatus as illustrated in this figure, 24 as with the other illustrated apparatus, a modification of the target, as indicated in FIG. 5, can be used. In 26 this figure the target, generally indicated as 80, is 27 shown to include a flat, rectangular plate 82 which is 28 clamped to the rim 84 of a wheel 86 by means of a 29 clamping mechanism 88. It will be evident from FIG. 5 that the plate 82 is angled relative to the primary 31 plane of the wheel 86 as well as to its axis of rotation 32 89.
33 The precise orientation of this plate 82 relative 34 to the wheel axis is, however, not known and will, as is described later, be computed with respect to the wheel 36 axis by the determination of a run-out factor for this 37 wheel. The general orientation of the plate 82 is, 38 however, chosen so that it can be adequately viewed by 39 the video camera 30 as it sights onto it.

WO 97'/14016 PCT/US96/16362 1 Finally, plate 82 includes a plurality of dots 90 2 which, as shown, constitute a pattern not unlike that on 3 the target illustrated in FIG. 3. , 4 With targets of this nature, the images formed by the video cameras 30, when used together with the 6 apparatus illustrated in FIG. 4, will be something like 7 that illustrated in FIG. 6. In this figure, it is 8 apparent that four discrete images 92, 94, 96, 98 are 9 formed to make up the complete image, generally indicated as ~99, formed by the video camera 30. Each of 11 the four images that make up the complete image 99 is an 12 image of one of the rectangular plates 82, respectively 13 disposed on the four wheels of the motor vehicle. For 14 example, the image 92 at the top of the picture 100 could correspond to the plate 82 on the right rear wheel 16 24R of the vehicle 20. Similarly, image 94 could 17 correspond to the right front wheel 22R, image 96 to the 18 left front wheel 22L and image 98 to the left rear wheel 19 24L.
The advantage of the target 80 when used with the 21 apparatus illustrated in FIG. 4 is that a single image 22 can be taken simultaneously of all four wheels. This 23 single image can then be processed, in very much the 24 same way as described above, to yield the orientation and location of all the wheels to each other. More 26 particularly, the relative orientation of the right 27 front wheel to the left front wheel and the right rear 28 wheel to the left rear wheel can be calculated.
29 On either end of the images 92, 94, 96, 98 a~pair of dots 100 can be seen. These dots 100 are in fact 31 images of the dots on the respective pan-and-tilt 32 mirrors referred to in the discussion of FIG. 2. As was 33 pointed out in that discussion, these dots are used to 34 calculate the orientation of the pan-and-tilt mirrors to the view path of the camera; a calculation which is 36 essential to determine both the orientation and the 37 location of the primary plane of each of the wheels of 38 the vehicle.

CA 02232534 1998-03-19 ~ , 1 In addition, this figure illustrates that the 2 images of the marks 100 can be separated from the images 3 of the patterns on the plate by means of a vertical line 4 101. This line 101 serves as a demarkation line between the pattern (from which the orientation of the target is-6 calculated) and the image of the dots 100 (from which 7 the orientation of the pan-and-tilt mirrors is 8 calculated).

Runout Factor Calculations 11 In FIG. 7 of the drawings, a method of calculating 12 the run-out factor for a target 104 mounted in a 13 slightly different way on a wheel 103 is illustrated.
14 In this method, the wheel 103 is slowly rotated while a number of different images of the target 104 are taken.
16 This target is, for the sake of clarity, off-set fairly 17 substantially from the center of the wheel. In 18 practice, however, the target may be mounted closer to 19 the center, much like the target illustrated in FIG. 5.
For each image, the inclination of the plane of the 21 target, as well as its location in space is calculated.
22 Once these have been determined for each image, they are 23 integrated to define a surface of revolution 106. This 24 surface of revolution 106 will represent the path which the target 104 tracks as the wheel is rotated about its 26 axis, and the axis of rotation 108 thereof is the same 27 as the axis of rotation of the wheel. This means that a 28 plane perpendicular to the axis of rotation 108 of the 29 surface of revolution 106 will be parallel to the primary plane of the wheel 106. As the surface of 31 revolution 106 is determined, its axis of rotation 108 32 is determined and, therefore, the orientation and 33 position in space of the primary plane of the wheel of 34 the vehicle can be determined.
From these results, the run-out factor can be 36 determined by calculating the angle between the plane of 37 the target and the primary plane of the wheel. This 38 run-out factor is then stored in the computer 32 and . , CA 02232534 1998-03-19 ' WO 97/14016 PCT/US96I16362 1 used when the alignment of the wheel is calculated from 2 a single image of the target.
3 The calculation of the run-out factor can also be 4 used to determine whether or not the suspension of the vehicle is badly worn. Using the method of the 6 invention an apparent run-out factor (i.e., the 7 orientation of the target with respect to the wheel) can 8 be determined for each image which is taken of the 9 target. From this group of individual run-out factors a mean value can be calculated (which will represent the 11 true "run-out" factor) as well as the extent of the 12 deviation from the mean of the individual factors. If 13 this deviation is above a certain tolerance, this 14 indicates that the suspension of the motor vehicle is worn so badly that it needs to be attended to.

17 Accuracy Determination 18 Turning once again to the targets, it should be 19 realized that an important feature of the target illustrated either in FIG. 3 or 5 (or any other target 21 for. that matter) is that it should have sufficient data 22 points to allow redundant calculations to be made using 23 different sets of data points. This will yield multiple 24 wheel alignment angles which can be averaged out to improve the accuracy of the final measurement. In 26 addition, a statistical distribution of the different 27 alignment angles calculated for each wheel can be used 28 as a measurement of accuracy of the operation of the 29 apparatus. If a suitable check is built into the computer 32, a statistical distribution such as this can 31 enable the computer 32 to determine whether or not 32 sufficient accuracies exist and, if not, to produce a 33 signal which can alert an operator to this fact.
34 Similarly, if the above checking indicates that one or more of the targets used yields) unacceptably poor 36 results while the remaining targets) yield acceptable 37 results, it can be assumed that some of the targets 38 being used are unacceptable. The computer can give an 39 indication to this effect and the operator can, for WO 97/14016 , PCT/US96/16362 1 example, be instructed to remove, clean or repair the 2 offending target (s) .
3 A further benefit derived from forming suitable 4 multiple images and computing a statistical analysis, is that the computer 32 can determine whether or not enough 6 images have been taken to suitably ensure the required 7 accuracy of the alignment measuring process. If 8 insufficient readings exist, the computer can direct the 9 apparatus to take further readings which, although sacrificing speed, would result in improved accuracy of 11 the measurement.
12 Furthermore, the target could include a machine-13 readable, e.g. a bar code or the like, which can be used 14 for identification, target tracking, intensity threshold measurement, evaluation of illumination quality, and 16 encoding of defects to allow the use of cheap targets.
17 For example, if the target was twisted and the amount of 18 twist was encoded in the bar code, then the computer 19 could compensate for the twist.
Another important feature of the target is that the 21 pattern thereon should allow very quick and accurate 22 location of the pattern to an accuracy approaching 23 substantially less than a camera pixel. To achieve this 24 the pattern should exhibit a high contrast and be of a configuration which allows the specific apparatus used 26 to achieve the required speed and accuracy. In one 27 embodiment, retro-reflective materials are used for the 28 dots, and a color that is absorptive of the particular 29 light used is chosen for the background.
This apparatus also allows for calibration, which 31 is important as all optical systems have some geometric 32 distortion. The total image area of the apparatus 33 could, for example, be calculated using a perfect target -34 and the result used to determine correction values that can be stored for use when operating the system in 36 alignment procedures.
37 The absolute accuracy of the apparatus can be 38 checked or calibrated by using a simple 2-sided flat 39 plate target which is placed so that the apparatus views . CA 02232534 1998-03-19 WO 9,7/14016 PCT/US96/16362 1 both sides simultaneously. As the plate is flat, the 2 net angle (relative alignment) between the two planes of 3 the target should be zero. If not, a suitable 4 correction factor can be stored in the computer.
Alternatively, two views of the same side of the target 6 taken from different angles could be used for this 7 purpose.

9 Mathematical Algorithms Used This section provides the mathematics necessary to 11 reduce measurements made by the video camera to wheel 12 positions in space using instantaneous measurement.

14 Assumptions The camera system can be defined to include two 16 .planes positioned arbitrarily (within reasonable 17 constraints of visibility) with respect to one another.
18 One is the image plane, which maps what is "seen" by the 19 camera and the other is the object plane, which contains three-dimensional, essentially point targets.
21 Based on this, the assumptions made are:
22 (l) the camera principal axis is normal to the 23 image plane (most cameras are built this 24 way);
(ii) there exists, at a known distance of f 26 (i.e. the imaging system's focal length 2~ -when set at infinity) from the image plane, 28 along the camera principal axis, a point 29 called the center of perspectivity (CP) such that the behavior of the camera is 31 that the image of a viewed point anywhere 32 in the camera's field of view is to project 33 it onto the image plane by moving it along 34 a line passing both through the viewed point in space and. the CP;
36 (iii) the origin of the coordinate system fixed 3~ in the image plane is located at the center 38 of perspectivity, with z unit vector 1 directed toward camera along its principal 2 axis; and ' 3 (iv) the units of the image plane measurements 4 are the same as those of the object plate measurements.
6 These assumptions are commonplace in the visual '7 sciences.

9 Overview For this configuration, mathematics can be provided 11 to determine the relative orientations and positions of w 12 the object and image planes.
13 This mathematics can be used in 2 ways:
14 (i) during calibration, to find the position of the image plane with respect to the 16 location of an object plane of known position of a calibration target; and 18 (ii) during the alignment process, to find the 19 position and orientation of the primary plane of the target mounted on the wheels 21 of the vehicle. It is essential in this 22 step that the known coordinate system is 23 fixed in space, and that it remains the 24 same for all four wheels of the car.
As has been described above, once the location of 26 the target planes on the wheels is known, by rotating 27 the wheels, the axis of rotation of the wheels can be 28 determined, and from there, the alignment of the wheels.

Main AJ.c~orithm 31 It should be noted that this main algorithm 32 presents no treatment of the various pan-and-tilt 33 mirrors; this is done later.
34 The main algorithm requires the following inputs:
(i) A list of points expressed in object plane 36 coordinates.
3~ °q~ = (x~. Y~) . j - 1. n / n > 4 ' WO (7/14016 PCT/US96116362 1 These are actually three-dimensional points, but the 2 object plane coordinate system can always be chosen so 3 that the third coordinate z; = 0. , 4 (ii) A corresponding list of image plane point coordinates 'q~ _ (u~, vj) , j - 1, n.
6 For these inputs, the algorithm produces an output 7 which is a homogeneous coordinate transform matrix 8 expressing the center of perspectivity and unit vectors 9 fixed with respect to the principal axes of the image plane. This matrix will normally be inverted, and then 11 applied to transform the viewed points into image system 12 coordinates.

14 Step 1: Determine a Collineation Convert all the two-dimensional input coordinates 16 to affine form and find a 3x3 transformation matrix T
17 such that:
(1) kiui x~
ki vt . T Yi ki 1 for i = 1, n and where the k; are arbitrary scalar 26 constants.

27 One way in which the transformation matrix T can be 28 determined is given below.

Step 2: Determine transforms and of kev points 31 invariants.

32 The transform matrix T will transform points in the 33 object plane to points in the image plane under the 34 projectivity whose center is the center perspectivity of CA 02232534 1998-03-19 , , 1 (CP). When inverted, it will also perform the reverse 2 transformation, viz:
(2) mixi aj .
m jyi _ T_~ vi mi 1 It will be noted that the whole equation may be 6 multiplied by an arbitrary scalar and still remain true.
7 The value m; is such a scalar, and is required to permit 8 normalization of (u; v; 1)T so that its third coordinate 9 is a unit. The matrix T is also useful for transforming lines, which are dual to points on the projective plane.
11 The equation of a line in the projective plane is:
(3) xi C1 C2 C3 , ~ x2 ~ = O C T X = O

17 Where c is the coordinate vector of the line and X
18 is the specimen vector. Any homogeneous representation 19 of a point which satisfies equation 3 lies on the line.
Suppose that an object co-ordinate °c lies on a line, 21 then:
(4) ° ~ ~ f °x l =
26 is the equation of a line in the object plane, expressed 27 in object plane coordinates. Using equation 2 we can 28 transform to image plane co-ordinates:
( ° c l T-i iX = 0 (S) or Therefore tC~~.tX~
(6) f.°1-IT°lf.°l is the way to transform line coordinates from the object plane to the image plane'and (7) fwl-f=-11 ~l 11 is the way to perform the inverse transformation.
12 Note that the projective plane differs from the 13 non-projective plane in that it includes points at 14 infinity whose projective coordinate is 0: These points 15 together constitute a line at infinity, whose 16 coordinates are [0,0,1] viz.
(8) x 0 0 1 ) yYy W = 0 26 This is illustrated in FIG. 8a which represents a side 27 view of an object plane OP and image plane IP positioned 28 non-parallel to each other at some angle A.
29 The object plane OP intersects a plane parallel to 30 the image plane IP but passing through the center of 31 perspectivity CP. This plane is called the view image 32 plane VIP and intersects the object plane OP at the 33 "vanishing line" mapped to the object plane, shown as 34 point VLO. Similarly, the figure shows a plane parallel 1 to the object plane called the viewed object plane VOP
2 which intersects the image plane IP at a "vanishing 3 line" mapped to the image plane, shown as point VLI.

CA 02232534 1998-03-19 , , 1 As VIP is parallel to IP they intersect at 2 infinity. The collineation matrix T can therefore be 3 used to map the line at infinity of the image to its 4 transformed position in the object plane as follows:.

1~.0 = b2 b3 1 (9) and likewise:
(l0) o VLI =- a2 - Z' 1 0 16 By the assumptions stated above with respect to the 17 camera system, the coordinates of the principal point of 18 the image PPI are: (11) PPI = 0 28 The coordinates of the principal point of the 29 object PPO are:
(12) PPO =_ c2 - Z,_1 0 c3 1 1 Step 3: Complete Remainincr Inclination Values 2 The minimum distance between a line in a projective 3 plane with line coordinates [z1 z2 z3] T and a point with 4 coordinates [pI pi Pa] T is given by (13) d 2iPi+2aPa+2sPs -_ ~ Ps 2i t 2z , , CA 02232534 1998-03-19 ' WO X7/14016 , PCTlL1S96/16362 .

This makes it possible to solve for DI, A and. DO:
(14) DI = a3 a~ + ai (15) A . - arctan DI
(16) DO = b1 c1 + bz cz + b3 c3 Cs bi + bz Step 4: Compute Pan Values FIG. 8b illustrates a plan view of the object plane, looking down from the center of perspectivity.
We have (17) arctan - ~ b zl (18) Let x+ = sgn (SX - PPO) . °x b1 Cl + b2 C2 + b3 C3 _ -s n ~ blC3 g Y+ = sgn (S3, - PPO) . °~, b1 cl + bz cz + b3 C3 g - -s n ~ bzc3 (19) Step 5: Solve for Remaining Unknowns Referring to FIGS. 8a and 8b together:
(20) DCP=ADO . sin0~

~ CA 02232534 1998-03-19 WO 97!14016 PCT/L1S96/16362 (21) x+ . sin 8 cos ~
° Z - y+ . sin 8 sin ~
'' cos 8 ° 0 f = ° PPI = ° PPD + DCP. ° Zi _,c!

_ Ca + DCP ° Zi (22) 1 This is the origin of the image plane coordinate system 2 expressed in object plane coordinates. It is located at 3 CP.
4 °~~ and °g! the remaining unit vectors can be S computed by transformation of the corresponding unit 6 vectors in the image plane, and subsequent 7 orthogonalization with respect to z!.
(23) Let xi 1 xa - T-1 0 x3 0 (24) xi _ c!
x3 C3 Then °~ - xa c2 - _ x3 c3 _ can be orthoaonalized with respect to °z.
(25) obi _ off' - (off'. oZ1) °Z1 WO ~7/140I6 PCT/US96/16362 and renormalized (26) °x~ = °~~ , o~ .

(27) Similarly, let 3'i 0 Y2 = T-~ 1 y3 . 0 (28) 5'i _ c'i Then ° W - -Ya _ ca (29) °zi) °z.
and (30) ° Y~ _ °
s Final 1y ( 31 ) X2 1 Y1 I Z1 1 ~Z
F _ ____.____1____.____ 0 ; 0 ; 0 ; 0 14 Frames transform from image space to object space 15 is the frame to return, and to express points given in 16 the object plane coordinates in terms of the coordinate 17 system fixed with respect to the image plane, we note (32) tFo = oF~i CA 02232534 1998-03-19 , , WO 97/14016 , PCT/US96J16362 .

and ( 3 3 ) xx sqk = sFo Yk 32 That is the general case, but there is also the 33 special case when the object and image planes are parallel. This is detectable when VLO or VLI (equations 1 9 or 10) turn out to lie at infinity themselves (meaning 2 their first two coordinates lie sufficiently close to 3 0) .
4 In this case, (34) J
9 and the distance DCP can be determined by taking any point in the object plane (xk , yk) whose corresponding 11 (ux , vk) is not zero and calculating according to the 12 diagram in FIG. 8 (c) (35) Let r° = xk + Yk (36) rZ = uk + vk (37) DCP = r° f rj 25 and then proceeding as from equation (22).
26 This concludes the description of the main 27 algorithm to determine plane displacements.

. CA 02232534 1998-03-19 ' WO X7/14016 PCT/US96/16362 1 9.4 Determination of Transform Matrix 2 This section illustrates how to calculate the 3 transform matrix T used in equation (1).
4 The method presented here is an analytic method which maps between only 4 coplanar points and is based 6 on the fundamental theorem of projective geometry which 1 ' tells us that given four points in the projective plane:
Pi - (xi Yi wi) (38) P2 - ( xz ' Yz wz ) p3 - ( x3 Y3 w3 ) pa - ( xa Y4 wa ) 6 constants c1, cz and c3 can be found such that . Pa - ciPi i -1- Czp2 + CgP3 ( 3 9 ) 8 When this is represented in matrix form:
(40) xi Yi y x4 Yq Wq ~ _ ~ C1 CZ C3 ~ X2 Y2 W2 then the matrix M consisting of (41) c1 0 0 x1 Yl wl M = 0 C2 0 x2 ya ta2 0 0 C3 x3 y3 Ws will transform the ideal points origin and unit points as follows:
(42) p1 = ( 1 0 0 ) M = ixM (unit x vector) pz = ( 0 1 0 ) M = i~,M (unit y vector) p3 = ( 0 0 1) M = oM (origin) p4 = (1 1 1)M = uM (unit point) Therefore, to construct a transform which maps four arbitrary points p1, pz. P3. pa to four arbitrary other points q1, qz, q3, q4, two transforms must be constructed:

M; M~ ' ( 4 3 ) ~-z "'~ P1 1z "'~ qi 1.y "~ P2. 1Y ~ q2 0--'P3 0-' a -' P4 a -~ qa 1 and then M, such that 2 q;=p;M (44) 3 is given by 4 M = M;1 MZ (45) Note that in this section, the p's and q's are now 6 vectors. In the main section, column vectors are used, 7 so 8 T = MT (46) 9 Finally, another method (not illustrated here) accepts more than four points and does a least-squares 11 approximation using pseudo-inverses. This second method 12 can be used in the case where the number of points 13 measured has been increased to compensate for expected 14 errors.
16 Allowance for Pan-and-Tilt Mirrors 17 After the imaged data points have been converted' 18 back to three-dimensional points given in image plane 19 coordinates, it remains to make allowance for reflections by the beam splitter assembly and the pan-21 and-tilt mirrors.
22 If 'x is a point to be reflected, and 'n is a unit-23 length normal to the plane of reflection, ixfl is a point 24 in the plane of reflection (all expressed in image-plane coordinates) then 'x" its reflection is given by (47) x__1 _ _1__ 2intnT_ ~ _ 2t31iI1ri3C _ sX
r ~j' o 1 0 i 1 1 27 The matrix above is a standard displacement style 28 transform which may be inverted using standard methods 29 though there is no need to do so in the present 3o application. These matrices may also be cascaded as 1 usual from right to left, to deal first with the beam-2 splitter and then with the pan-and-tilt mirror, but the 3 reflection plane point 'xfl and normal 'n for the pan-and-4 tilt mirror must first be transformed by the beam- -splitter reflection matrix before the pan-and-tilt 1 mirror reflection matrix is formed from them.
2 Finally, it should be noted that when the main 3 algorithm is used to find the position of the pan-and-4 tilt mirror, once these have been reflected through the 5 beam splitter. 'zo and '0o are directly usable as normal 6 and point in the reflection plane directly.
7 A subsequent use of an iterative fitting procedure 8 may result in improved accuracies. -Other mathematical processes can also be used to l0 process the images detected using.the apparatus of the 11 present invention.

13 Alternative 'I~o-Camera Embodiment 14 In Fig. 9 of the drawing, an alternative embodiment of the present invention utilizing a pair of fixed, 16 spaced-apart cameras is depicted at 110. A four-wheeled 17 vehicle positioned on a lift ramp 111 for wheel 18 alignment is suggested by the four wheels 112, 113, 114, 19 and 115. In the usual case, the rack 111 will include pivot plates (not shown) to facilitate direction change 21 of at least the front wheels. In this embodiment a 22 camera supporting suprastructure includes a horizontally 23 extending beam 116 affixed to a cabinet 117. The 24 cabinet 117 may include a plurality of drawers 118 for containing tools, manuals, parts, etc., and may also 26 form a support for a video monitor 119 and input 27 keyboard 120.
28 Mounted at each end of the beam 116 is a camera and 29 light source subsystem respectively designated 122 and 124. The length of beam 116 is chosen so as to be long 31 enough to position the camera/light subsystems outboard 32 of the sides of any vehicle to be aligned by the system.
33 The beam and camera/light subsystems 122, 124 are 34 positioned high enough above the shop floor 125 to CA 02232534 1998-03-19 , , WO 97/I4016 PCTiUS96/16362 1 ensure that the two targets on the left side of the 2 vehicle are both within the field of view of camera 3 assembly 122, and the two targets on the right side of 4 the vehicle are both within the field of view of camera assembly 124. In other words, the cameras are 6 positioned high enough that their line of view of a rear 7 target is over the top of a front target. This can, of 8 course, also be accomplished by choosing the length of 9 beam 116 such that the cameras are outside of the front targets and have a clear view of the rear targets'.
11 Details of the camera/light subsystems 122, 124 are 12 discussed below with respect to Fig. 10.
13 , In accordance with this embodiment, a target device 14 126, including a rim-clamp apparatus 128 and a target object 130, is attached to each wheel. A suitable rim-16 clamp mechanism is discussed in U.S. Patent No. ' 17 5,024,001 entitled "Wheel Alignment Rim Clamp Claw". As 18 will be described in more detail below, the preferred 19 target object has at least one planar, light-reflective surface with a plurality of visually perceptible, 21 geometrically configured, retro-reflective target 22 elements 132 formed thereon. Such target surfaces may 23 be formed on one or more sides of the target object. In 24 use, each target must be positioned on a vehicle wheel with an orientation such that the target elements are 26 within the field of view of at least one of the 27 camera/light subsystems.
28 In Fig. 10 of the drawing, further detail of the 29 ' camera and lighting components is illustrated. Mounted within the partially broken-away end of beam 120, the 31 subsystem 122 is shown to include a lighting unit 140, 32 comprised of a plurality of light emitting diode (LED) 33 light sources 142 arrayed about an aperture 144 through 34 which the input optics 146 of a suitable video camera 148 is projected. The light array in the preferred 36 embodiment includes 64 LEDs (a lesser number being shown 37 for simplicity of illustration) which provide a high-38 intensity source of on-axis illumination surrounding the 39 camera lens, to ensure that maximum light is retro-reflected from the targets. In order to discriminate against other possible sources of light input to the camera 148, a narrow band filter matched to the light spectrum of the LEDs may be positioned in front of the lens 146.
Although any suitable type of video camera can be utilized, in accordance with the preferred embodiment a CCD device, such as that manufactured by Phillips is utilized. This camera has a resolving power suitable for the present application.
In Fig. 11, an example of a target in accordance with a preferred embodiment is depicted and includes a plurality of light-reflective, circular target elements or dots of light-colored or white retro-reflective material disposed in an array over a less reflective or dark-colored surface of a rigid substrate. Suitable retro-reflective materials include NikkaliteT''t 1053 sold by Nippon Carbide Industries USA, Scotchlite~ 7610 sold by 3M Company and D66-l5xx~ sold by Reflexite, Inc.
The target includes multiple circular dots so as to ensure that sufficient data input may be grabbed by the camera even in the case that several of the target elements have been smudged by handling or are otherwise not fully detectable. In accordance with the preferred embodiment a well defined target includes approximately 30 circular dots very accurately positioned (within 0.0002") with respect to each other. By way of specific example, the target illustrated in Fig. 11 might include 28 circular dots of 1" diameter very accurately positioned on a 2" x 2" grid, with four 11/" dots and a single 11/" diameter dot strategically positioned within the array. By mathematically moving the mathematical image of a target until the mathematical position and orientation of the dots line up with the dots of the real target in the real image, position and orientation information can be obtained. This mathematical manipulation of a well defined target until it is oriented the same way as the image is called "fitting the target". Once the fitting is accomplished, the WO 97/14016 PCTlLTS96/16362 1 position and orientation of the target is very 2 accurately known (to within 0.05" and 0.005°). Such 3 accuracy is obtainable because the target is made to 4 very strict tolerances and because the design enables.
measurement of many points (1,500 measured points, i.e.
6 30 or so fiducials (dots) each with detected 50 edge 7 points). Furthermore, the use of subpixel interpolation 8 enhances the accuracy of measurement to beyond the pixel 9 resolution of the CCD cameras.
The target is typically manufactured using a 11 photolithographic process to define the dot boundaries 12 and ensure sharp-edge transition between light and dark 13 areas, as well as accurate and repeatable positioning of 14 the several target elements on the target face. The target face may also be covered with a glass or other 16 ,protective layer. Note that since all information 17 obtained from a particular target is unique to that 18 target, the several targets used to align a vehicle need 19 not be identical and can in fact be of different makeup and size. For example, it is convenient to use larger 21 rear targets to compensate for the difference in 22 ~ distance to the camera.
23 In order to accurately determine the position 24 between the wheels on one side of the vehicle and the wheels on the other side of the vehicle, the system must 26 know where one camera is positioned with respect to the 27 other camera. This is accomplished during a calibration 28 and set up operation wherein, as depicted in Fig. 9, a 29 larger target 150 (presently 3' x 3') is positioned in the field of view of both cameras, typically along the ' 31 centerline of the rack 111, and the approximately 30 32 feet away from the cameras. Information obtained from 33 each camera is then used to determine the relative 34 positions and orientations of the cameras. More specifically, since each camera will indicate where the 36 target is with respect to itself, and since each is 37 viewing the same target, the system can calculate where 38 each camera is located and oriented with respect to the 39 other. This is called a relative camera position (RCP) ~ ~ - CA 02232534 1998-03-19 ' WO 9'7/14016 PCT/US96/16362 1 calibration. Such calibration allows the results 2 obtained from one side of the vehicle to be compared to 3 the other. Thus, by mounting the two cameras rigidly 4 with respect to each other and them performing an RCP.
calibration, the system can be used to locate the wheels 6 on one side of the vehicle with respect to the other 7 side from that point on. This is to say that the RCP
8 transfer function is used to convert one camera's 9 coordinate system into the other camera's coordinate system so that a target viewed by one camera can be 11 directly related to a target viewed by the other camera.
12 The inspection process of the present invention is 13 monocular, meaning that by using one camera in one 14 position, the position and orientation of a target with respect to the camera can be determined. This, of 16 course, requires that the target be in view of the 17 camera to accomplish the measurement. But since one 18 camera can only conveniently view one side of the 19 vehicle at a time without using reflectors as earlier described above, two spatially related cameras must be 21 used to view both sides. The RCP transfer function then 22 allows the information obtained by the two cameras to be 23 coordinated and have the same effect as if all of the 24 information had been obtained by a single camera. An advantage of the use of such a system is that, since 26 each wheel is independently inspected and merely related 27 back to the others, the system is independent of level 28 and does not require leveling of the vehicle support 29 rack or floor. Moreover, it is not necessary that the axles of all wheels be at the same~height, i.e., 31 differences in tire sizes or inflation will not 32 adversely affect measurement.
33 In operation, once the system has been calibrated 34 using the calibration target 150 as illustrated in Fig.
9, a vehicle may be driven onto the rack 133, and, if 36 desired, the vehicle lifted to an appropriate repair 37 elevation. The target assemblies 126 are then attached 38 to the wheel rims and manually oriented so that the 39 target surfaces face the respective camera/light WO 97/14016 PCTlUS96/16362 1 subsystems. The vehicle and model year are then entered 2 into the keyboard 120 along with other relevant 3 information which may include the vehicle VIN number,, 4 license number, owner name, etc. The system database includes specifications for each model that might be 6 inspected, and upon identification of the particular -vehicle under inspection extracts such information to 8 assist in quickly locating the target images.
9 Alternatively, previous inspection history can be used l0 to indicate likely target location.
11 The targets are highly accurate and their position 12 and orientation relative to the rim of the wheel to 13 which they are attached is known to an accuracy of 0.01"
14 and 0.001°. If each wheel was perfect and the clamp was perfectly mounted one could argue that the wheel axle 16 would be normal (90° in all directions) to the wheel 17 plane determined by the rim edge. However, since wheels 18 are normally not perfect and targets are not always 19 perfectly mounted, such information would only indicate orientation and position of the wheel plane and not 21 necessarily provide accurate information as to the 22 orientation of the wheel axis. Such assumption is thus 23 not made. However, by rolling the wheel from one 24 position to another a new image can be taken, and from the position and orientation of the target in the two 26 positions, the actual position and orientation of the 27 wheel axis can be calculated. .
28 Similarly, to calculate the steering axle (about 29 which the wheels turn when the steering wheel is turned) two target positions are again compared, one with the 31 wheels turned to one side and one with the wheels turned 32 to the other side. Calculation of the axis about which 33 the targets must have been moved thus identifies the 34 position and orientation of the steering axis.
Now knowing where each wheel axle is located and 36 how it is oriented, where the steering axles are located 37 and how they are oriented, the vehicle can be 38 mathematically modeled in three dimensions, and the 39 alignment values in toe, camber, caster, thrust angle, ' W0~97/I40I6 PCT/US96/16362 1 etc. can be displayed with respect to the vehicle 2 itself .
3 Once the targets are installed on each wheel and , 4 the system is energized, enough information is available to generate an image such as that depicted in Fig. 2a.
6 However, as pointed out above, because the rotational axis of the wheels may not be exactly normal to the 8 wheel plane as defined by the outside perimeter of the 9 rim to which the target assembly is attached, the system operator will be instructed to move the vehicle forward 11 or aft 6 or 8 inches so as to rotate the wheels through 12 about 30° of rotation. With measurements taken of at 13 least two different wheel positions, the system can 14 optically obtain enough information to accurately determine true axle position and orientation for each 16 wheel. Highly accurate computations can then be made 17 and displayed on an updated screen, as depicted in Fig.
18 2a.
19 At this point, the actual operator alignment procedure can proceed, and since the inspection is 21 continuous, the results of each adjustment will be 22 reflected on the system video screen. In the preferred 23 embodiment of the present invention, the operator can 24 select various levels of assistance, including actual depictions of the location and parts to be adjusted to 26 provide corrective action. Such information can even 27 include the appropriate choice of tool to be used.
28 As pointed out above, since each camera is 29 referenced to the other, it is not necessary that the supporting rack be level or even that all wheels lie 31 within the same plane. However, although each wheel 32 inspection is independent of the others, a reference 33 plane must be identified. This can be accomplished by 34 defining a reference plane that passes through the axles. But since one of the axles may not lie in the 36 plane defined by the other three, some liberties 'must be 37 taken. For example, for the purpose of aligning the 38 front wheels, one might use the plane defined by the 39 front axles and the average of the rear axles. A

1 similar procedure might be used with respect to the rear 2 wheels, etc. Wheel alignment would then be referenced '3 to such a plane or planes. In addition, wheel position 4 and thrust line measurements would also be referenced to such a plane or planes. Moreover, because of the 6 independence of measurement, once a reference plane is defined, should one of the targets be blocked from view 8 or become loose or even dislodged from a wheel, it will 9 not necessarily affect measurements associated with other wheels.
11 Having now described several embodiments of the 12 present invention suitable for use in aligning the 13 wheels of a vehicle, and having pointed out that the 14 position and orientation of each target and associated wheel may be determined independently of the other 16 targets (and wheels), it will be appreciated by those 17 skilled in the art that by modifying the target 18 attachment structure to enable the targets to be affixed 19 to other particular points on the vehicle, or to another type of structure, such as, for example, a building 21 structure, an article of manufacture, a robot arm, or 22 even territorial space, the same system can be used to 23 measure relative spatial location or alignment of the 24 several points to which the targets are affixed. For example, in the case of an automotive vehicle or the 26 like, one might use the described. system to measure 27 vehicle chassis or body alignment, or perhaps ride 28 height. And because the data is updated at a high rate, 29 "jounce" measurements (i.e., a measurement of suspension dynamics) can be made. In the case of articles of 31 manufacture, one might wish to embody a target in the 32 form of a label and affix the label to parts on an 33 assembly line, and then use the present invention to 34 track the position and/or orientation of the article as it moves down the line. In the case of a robot arm, one 36 or more targets affixed to various moving parts could be 37 used to accurately follow the motion of the arm as 38 objects are carried thereby. In the case of building 39 structures, one might use a system in accordance with ~ ~ CA 02232534 1998-03-19 ' WO 87/14016 PCT/US96/16362 1 the present invention to determine or maintain alignment 2 of various points on the structure relative to other 3 points. In the case of territorial space, one might use , 4 the system to develop topological surveys of ground surface contours.
It will also be apparent that more than two cameras 7 could be used to inspect objects or fields of view not 8 readily inspectable with one or two cameras. In such 9 case an RCP transfer function calibration procedure l0 similar to that described above would be followed.

12 Additional Features of the Invention 13 .As indicated above, this invention can also be used 14 to determine the condition of the shock absorbers of the vehicle. This is done by firstly ~~jouncing~~ the 16 vehicle. Jouncing a vehicle is a normal step in 1~ alignment procedures, or, for that matter, checking the 18 shock absorbers, and entails applying a single vertical 19 force to the vehicle by, eg. pushing down onto the hood of the vehicle and releasing the vehicle, to cause it to 21 oscillate up and down. Secondly, as the vehicle 22 oscillates up and down, the apparatus of the invention 23 takes readings of the targets on each of the wheels. In 24 so doing, the movement of the targets, which will define a dampened waveform, can be monitored to determine the 26 extent of the dampening. If the dampening is not 2~ sufficient (i.e. the up and down movement'or rocking of 28 the vehicle does not stop soon enough) this indicates 29 that the shock absorbers are faulty.
This method is particularly advantageous in that a 31 determination can be made as to the soundness of a 32 specific shock absorber; a result which can be indicated 33 to the operator of the alignment apparatus by means of 34 the computer 32.
It will be evident that in the determination of the 36 condition of the shock absorbers of the vehicle, any 3~ suitable portion of the body of the motor vehicle can be 38 selected to monitor the oscillation of the vehicle. So, 39 for example, the apparatus can focus on the edge of the 1 wheel housing or, alternatively, a small target placed 2 on a convenient position on the body work of the motor 3 vehicle.
4 In addition, this apparatus can be used to calculate the ride height of the motor vehicle. This 6 parameter is particularly important in the determination of the alignment of the wheels of vehicles such as pick 8 ups which, in operation, may carry a load. This load 9 would have the effect of~lowering the vehicle and it is, therefore, preferable to make allowance for this during 11 alignment procedures. Traditionally, the ride height, 12 or height of the chassis of the vehicle from the floor, 13 is determined by physically measuring it with an 14 instrument such as a tape measure. This measurement is then compared to standard tables which yield a 16 compensation factor for the vehicle concerned.
17 The method and apparatus of this invention can, 18 however, make this measurement directly by viewing an 19 appropriate portion of the body and determining its height from the test bed on which the vehicle rests.
21 Once this height has been determined it can be compared 22 to standard look-up tables stored within the computer 23 which can, in turn, produce the compensation factor.

Advantactes of the Invention 26 A general advantage of the apparatus of this 27 invention is that it is relatively simple to use as no 28 delicate mechanical or electronic equipment need be 29 attached to the wheels of the motor vehicle concerned.
As the sensitive and delicate equipment is mounted 31 within a housing which stands independent and distant 32 from the motor vehicle being tested, no damage can be 33 caused to it if the motor vehicle were, for example, to 34 be driven off the wheel guides. Whereas prior art heads can be knocked out of calibration by simple jarring or 36 dropping, it takes major damage to the wheel-mounted 37 components to affect the calculated results.
38 Another advantage is that the equipment requires 39 very few operator commands and could readily be made 1 hands free with simple auditory outputs and equally 2 simple voice recognition means to receive and/or record 3 operator responses and/or commands.
4 The present invention has the further advantage that alignment determinations can be done relatively' 6 quickly. This allows a higher turn around rate within 7 the business conducting the alignment determinations.
Still further advantages of this apparatus is that 9 it can be placed, as is illustrated in FIG. 4, above and out of the way of the motor vehicle being tested. This 11 has the distinct advantage that the chances of damaging 12 the sensitive alignment determining apparatus is 13 substantially reduced as the apparatus is out of the way 14 of the motor vehicle. Another advantage of this configuration is that the measuring apparatus uses 16 minimal floor space and has no equipment blocking access 17 to the front of the motor vehicle.
18 Furthermore, as the vehicle can be backed up and 19 driven forward, this apparatus has the advantage that it is unnecessary to jack the vehicle up to make the 21 required calculations for "run-out". In addition, this 22 apparatus can be used to determine information other 23 than the relative alignment of the wheels. For example, 24 the alignment apparatus, if equipped with a suitable character recognition capability, could be used to read 26 the license plate of the motor vehicle which could, in 27 turn, yield information such as the make and model of 28 the vehicle and its service history (if available) and, 29 therefore, the required alignment parameters of such vehicle. This would save the operator from entering the 31 motor vehicle's details into the apparatus. As more 32 manufacturers are adding bar codes to the VIN number 33 plates, similar information can also be obtained by 34 optically viewing and processing the bar-coded plate.
In addition, it would also be possible to optically 36 identify the vehicle type by comparing certain features 37 of the body or trim thereof to database information.
38 Yet another advantage of the invention is that no 39 wires, cords or beams of light pass in front of the ' WO 97/14016 PCT/US96/16362 1 vehicle being tested. As most alignment correction is 2 made by accessing the wheels of the car from the front, 3 wires, cords or beams passing in front of the vehicle 4 tend to get in the way of the technician. Often these wires, cords or beams are sensitive to being interfered 6 with and so their absence makes alignment correction 7 work much easier.
8 Related to this advantage is the fact that there 9 are no cords or wires passing between the targets on the wheels, nor are there any wires supplying power to the 11 targets from a remote power source. This absence of 12 wires or cords once again makes work on the vehicle 13 easier. -14 In addition, as the targets are not interlinked or interdependent, after initial capture of target images, 16 it is possible to block off one of the targets from the 17 camera's view without interfering with the orientation 18 calculations for the other wheels. In the prior art 19 devices described earlier, all the test heads are interdependent and cannot function if one of the heads 21 is "blocked" out.
22 It will be evident to those skilled in the art that 23 the concept of this invention can be applied in many 24 different ways to determine the alignment of the wheels of a motor vehicle. So, for example, the apparatus 26 could define a reference point for each wheel with the 27 referent point being located at, say, the intersection 28 of the axis of rotation of the wheel, with that wheel.
29 These points can then be processed to define an approximately horizontal reference plane, relative to 31 which the alignment of the wheels can be calculated.
32 This method has the particular advantage that the 33 rack on which the vehicle is being supported does not 34 have to be levelled, a process which requires expensive apparatus and which is necessary to define a horizontal 36 reference plane and which is used in prior art alignment 37 devices.
38 While the invention has been particularly shown and 39 described with reference to certain preferred ~WO 97/14016 PCT/LTS96/16362 1 embodiments, it will be understood by those skilled in 2 the art that various alterations and modifications in 3 form and in detail may be made therein. Accordingly, it 4 is intended that the following claims be interpreted as covering all such alterations and modifications as may 6 fall within the true spirit and scope of the invention.

Claims (16)

WHAT IS CLAIMED IS:

1. An apparatus for determining the alignment of motor, vehicle wheels comprising:
target means including at least first and second target objects for attachment to wheels on respective first and second sides of a vehicle under inspection, each said target object including a plurality of visually perceptible, geometrically configured target elements having known geometric characteristics and positional interrelationships;
optical inspection means defining a spatial reference system and including first and second cameras respectively forming first and second viewing paths intersecting said first and second target objects when they are attached to wheels of a vehicle under inspection, each said camera being operable to inspect an image of a corresponding target object as viewed along its corresponding viewing path and to generate image information describing the geometric characteristics and positional interrelationships of the imaged target elements;
processing means for relating such image information to predetermined reference information describing the known geometric characteristics and positional interrelationships of said target elements and for determining the position and angular orientation of said first and second target objects relative to said spatial reference system and generating first and second position and orientation information commensurate therewith; and means responsive to said first and second position and orientation information and operative to indicate the position and alignment of the wheels to which said first and second targets are attached.

2. An apparatus as recited in claim 1 wherein said means responsive includes a display means which uses said orientation information to indicate the alignment of said first and second wheels.

3. An apparatus as recited in claim 2 wherein the alignment of each said wheel is expressed in terms of caster, camber and toe measurements.

4. An apparatus as recited in claim 1 wherein said target means further includes third and fourth target objects for attachment to third and fourth wheels, respectively disposed on said first and second sides of the vehicle under inspection, said target objects each including a plurality of visually perceptible, geometrically configured target elements having known geometric characteristics and positional inter-relationships, said first and second camera means being respectively operative to establish third and fourth viewing paths intersecting said third and fourth target objects, and to inspect images thereof and generate image information describing the geometric characteristics and positional relationships of the target elements of each image, said processing means being further operative to relate said image information to predetermined reference information describing the known geometric characteristics and positional interrelationships of said third and fourth target elements to determine the angular orientation of said third and fourth target elements and to generate third and fourth positions and orientation information commensurate therewith, said means responsive being further responsive to said third and fourth position and orientation information and operative to indicate the alignment of the wheels to which said third and fourth target objects are attached.

5. An apparatus as recited in claim 4 wherein said processing means further determines the relative positions of said first, second, third and fourth wheels.

6. An apparatus as recited in claim 1 wherein said optical inspection means further includes lighting means associated with each camera means to provide on-axis target illumination therefor.

7. An apparatus as recited in claim 6 wherein said lighting means includes an array of light-emitting diodes arrayed around the optical axis of each said corresponding camera means.

8. An apparatus as recited in claim 7 wherein said light-emitting diodes are selected to have a particular wavelength characteristic and wherein wavelength-selective filter means are used to discriminate between diode light reflected from said target objects and light emanating from other sources.

9. An apparatus as recited in claim 1 wherein each said target object includes a planar plate having said plurality of target elements formed on a surface thereof and means for attaching the plate to a wheel of a vehicle under inspection.

10. An apparatus as recited in claim 9 wherein said means for attaching orients the plane of said surface substantially normal to the rim plane of the wheel to which the target object is attached.

11. An apparatus as recited in claims 1, 4, 6 or 7, wherein said target elements are formed of retro-reflective material.

12. An apparatus as recited in claims 1 or 4 wherein said target means includes clamps for clampingly engaging the rims of wheels.

13. An apparatus as recited in claim 1 and further comprising support means for supporting said first and second cameras at a predetermined separation from each other greater than the width of a vehicle having wheels to be aligned.

14. An apparatus as recited in claim 13 wherein said support means holds said first and second cameras in fixed positions at a predetermined elevation.

15. A method of determining the relative orientation of the wheels of a vehicle comprising the steps of:
establishing a first target means on a first wheel on a first side of said vehicle and establishing a second target means on a second wheel on a second side of said vehicle, each said target means including a plurality of target elements of known geometric characteristics and positional interrelationships, said first and second target means having a predetermined positional relationship to the wheels to which they are attached;
using a first camera means having a known spatial position and directional orientation to view said first target and to form a first detected image thereof, and using a second camera means having a known spatial position and directional orientation to view said second target to form a second detected image thereof;
determining the geometric characteristics and positional interrelationships of the target elements of each of said first and second detected images;
relating the determined geometric characteristics and positional interrelationships of said first detected image to the known geometric characteristics and positional inter-relationships of corresponding elements of said first target means to determine the angular orientation of said first target means;
relating the determined geometric characteristics and positional interrelationships of said second detected image to the known geometric characteristics and positional inter-relationships of corresponding elements of said second target means to determine the angular orientation of said second target means; and using said angular orientations to determine the alignment of said first and second wheels.

16. A method as recited in claim 15 and further including the step of relating the determined geometric characteristics and interrelationships of said first and second detected images to the known geometric characteristics and interrelationships of corresponding elements of said first and second target means to determine the positions of the axles of said first and second wheels.