CA1127259A - Method and device for inspecting a moving sheet material for streaklike defects - Google Patents
Method and device for inspecting a moving sheet material for streaklike defectsInfo
- Publication number
- CA1127259A CA1127259A CA336,470A CA336470A CA1127259A CA 1127259 A CA1127259 A CA 1127259A CA 336470 A CA336470 A CA 336470A CA 1127259 A CA1127259 A CA 1127259A
- Authority
- CA
- Canada
- Prior art keywords
- signal
- signals
- scanning
- sheet material
- measurement signals
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/89—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles
- G01N21/892—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles characterised by the flaw, defect or object feature examined
- G01N21/8921—Streaks
Landscapes
- Engineering & Computer Science (AREA)
- Textile Engineering (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
Abstract Identification of subtle streaklike defects in running webs and sheets by transversely scanning adjacent lateral sections of the webs or sheets by means of radiant energy means. Transmitted or reflected radiation is received on corresponding photocells that produce measurement signals that are representative for a defect in the webs or sheets.
The signal-to-noise ratio of the measurement signals is increased by multiplicative correlation, on the condition that ratio of the signal component Vd to the effective noise component Vn of the measurement signal Vm is greater than one.
The signal-to-noise ratio of the measurement signals is increased by multiplicative correlation, on the condition that ratio of the signal component Vd to the effective noise component Vn of the measurement signal Vm is greater than one.
Description
.z~
A ~--V. L~
~ his invention relates to a method and a device for inspecting a moving sheet material for streaklike defects, comprising means for directing radiant energy on one sur-i face of the sheet material and photocell means for measu-ring at distinct lateral positions of the sheet material ! the intensity of the radia-tion after transmission of -the - radiation through, or after reflection on the surface of the sheet material.
Enown devices operate satisfactorily for the detection of rather ooarse speck- and streaklike defects in moving sheets since in such cases the signal-to-noise ratio of the output sig~als of the photocells is sufficien-tly great to enable the signals, after their amplification, to be com-pared with a predetermined reference signal in order to identify a defect. Suchlike defect identifying sig~als ma~ either be stored in digital or analogue form in a memory together with the exact address of the defect, or they may directly be signalled as in the coating of webs in order to arrest or to re-adjust the coating, or they ma~ ~directly control a sorting mechanism as in the inspec-tion of sheets cut from a coated web. As examples of rather coarse defects can be mentioned specklike defects caused by -the inclusion o~ a stra~ge object, e~g. a dust particle in a coated layer, or a dry spot, e~g~ an air GV~10~1 ~.
- -', .~ .: ' ~_~ 2~f~
-- 2 --bubble included duri~g the coa-ti~g of a layer on a support.
In those cases, however7 where very minute defects have to be identified, fluttering of the material at the place of measurement a~d ~oise of the electronic circui-t~ may cause variatio~s in the uutput signals o~ -the photocells that are of the same order of magnitudo as those caused by a sheet defect.
I-t is known to use electronic correlatio~ techniques for improving the signal-to-~oise ratio, hereinafter re-ferred to as S/~ ratio, of the defect identifying signals.~ basic condition for the operation of such techniques is that the defect to be identified should be recur~ive during a number of, say at least 5, successive scans.
~his means that the defects have to ke of protracted dura-tion in the machine direction, that îs in the coating direc-tio~ in the case of the inspection of coated la~ers. In other words, only streaklike defect signals are suitable for ~/N ratio improvement.
The use of additive correlation in apparatus for the inspection of movi~g webs is disclosed for instance in G~-P
1,471,316 filed June 4, 1974 b~ Eastman Eod~k Co. and in U~P
~005,281 of Mark ~dwi Faulhaber a~d Edmund Haislett Smith Jr.
issued Januar~ 25, 19770 T~e mentioned use does not provide results that are e~tirely satisfactory for the following reasons~
~he ~/N ratio improvement that can be obtained is rather limited as will be further explai~ed in detail in the description of the invention. ~or instance, for a gain in the feedback loop of a delay line equal to 0.907 the S/~ gain is 12.78 db whereas for A = 0.99, the gain is 22~98 d~ in accordance with the formula : S/~ (in db) =
~0 log ~
Further, ~hereas it is possible to use for a factor = 0.90 an analogue circuitry7 the use of a factor A =
0~99 inevitabl~ ~ecessitates the use of a digital circuitry ~. Z~7~25~ .
in order to obtain a sufficient accuracy, and suchlike digi-tal arrangement requires a rather ex-tensive hardware layout.
Finally, the time t o:E the correlation clrcuit f'or reaching a significan-t S/~ ratio improvement is rathex important so tha-t in some cases i-t may ta'ke up -to ~00 seconds and more before -the correlated signal is availa-ble. It is the objec-t of the invention -to provide a new correlation technique in the inspection of a moving sh0et material for streaklike defects, that does not show the shortcomings menticned hereinbefore. ~he invention aims in particular at eonsiderably improving the S/~ ratio of photocell signals that are encountered in the identifica-tion of very subtle streaklike defects in moving sheet materials.
As an example of very subtle streaklike defects~
streaklike irregularities may be mentioned that occur at 'the coating on a web of a layer,,or of a combination of layers as the case may be, by means of a so~called cascade 20~ or'slide hopper coater. In certain applieation fields~
notably in photography, very stringent requirements are put to the uniformity of a coated layer, and it is shown that defect signals resulting from streaklike thickness 'deviations of an order of magnitude under 0.1 %, even when 25, additively correlated, are not discernable from the noise eomponent of the signals.
' ' As to the oceurrence of the mentioned defect in the said coating teehnigue, it is assumed that the most impor-,tant~eause for such defect is the passage of a web splice through the coa-ting bead. It appears that a web splice , entrains,air in the coating bead and that such air may re-main there for quite long periods, in some circums-tances eovering hundreds of meters of material, so that the coa-t-ing thickness at tha-t area is correspondingly dis-turbed.
G~.1031 :
~-~z~
_ 4 Considering the fact that in the manufacture of delicate materials, such as for instance radiographic film or film for graphic reproduction techniques, thickness deviations in coated layer in the order of magnitude of 0.1 % and preferably even less than 0.1 %
should be indentified, it will be understood that the signal--to-nolse characteristics of common detection systems do not enable a rel~able and rapid retrieval of the mentioned defect.
In accordance with the present invention~ a method for ;nspect;ng at least one longitudinal section of a moving sheet material for lo streaklike defects, by repeatedly transversely scanning said at least one section of the material by means of radiant energy capable of being modulated by said sheet material, and by receiving said modulated energy on at least one photocell thereby to detect at said at least one photocell signals that are periodically recurring in response with the scanning of the material and the variation of which signals is indicative of the presence of said defects, comprises improving the S/N ratio of those said periodically recurring signals the S/N ratio of which is greater than one, by delaying each measurement signal Vm over a time delay t corresponding with the scanning period P, multiplying a next signal Vm with a factor that is proportional to the product of said delayed signal and a factor A, delaying said multiplied signal over said time delay t and multiplying the next signal Vm with a factor that is proportional to the product of said previous multiplied signal and said factor A~
and so on for a number of times, wherein said factor A is at least equal to one, and wherein further Vm.A ~ 13 for said signal Vm being expressed in volts.
The notion "streaklike defect in a moving sheet material" stands in the present statement for a line-wise defect that runs parallel with the direction of movement of the sheet material during its inspection.
GV.1031 .{:r ~
-The term "sheet material" stands for webs that may have a length up to many hundreds of meters~ as well as for a plurality of distinct sheets that may have been cut from such webs and -that may be inspec-ted in succe~sion.
The term "modulated by" indicates that radia-~ion is attenuated on its pa-th towards the photocell, either b~
transmission through the sheet material or by reflectior from the surface of said sheet material~ ~aid defects modify -the attenuation.
~he expression "signals wi-th a ~/~ ratio greater than one" means in the present specification signals the defect-component of which is greater than the effective noise component. ~his means that peak-to-peak noise levels may be presen-t in the photocell signal that are much greater than the useful signal level but the effective value of which, that is the rms value determined over a time that is equal to the quo-tient of -the scanning time t by the number of discrete measurements during said time~
~he scanning period "P" is the time comprised between two successi~e scans of a given zone of said at leas-t one longitudinal section from edge to edge in one direction.
~ he measurement signal Vm may be the output signal of a photocell that must be compared with a given reference signal in order to establish the occasional deviation of said measurement signal, thereby to detect a defect at the lateral position of the sheet material inspec-ted by the photocell, but said signal Vm may also be~ and in a different embodiment preferably is, the difference, amplified as the case may be, between two successive signals of a mûltiplicity of sequential output signals that re~
present the scanning of the sheet material.
~ he term "scanning" stands for the repea-ted inspection of the ~heet material in a direction that runs -transverse with respect to the direction of movement of the sheet ma G~.10~1 ~2,7~5~
terial. ~he scanning may occur by means o~ a beam of ra-diatio~ that is projected on the sheet ma-terial and that is swept over the material. ~his sweeping may be done by means of a source of radiation mounted on a frame arranged for oscillation transverse wi-th regpect to -the m~-teIial, but said sweeping may also be carried out by mearls o~' a beam of radiation that is de~lec-ted by a rotati~g or pivo-tir~g mirror or -the like. Alterna-tively, ~he scannin~ may occur b~ means o~ a plurality o~ dis-tinct small light sources, for insta~ce light emitting diodes (~D's) that are mounted closely adjacent to each other in a row and that may be energized in sequence to produce a radia-tion spot that scans the sheet material.
~ he photocell may be arra~ged for lateral displacement sy~chronous with the scanning beam of radiation, but the photocell may also collect, for instance by means of a collecting glass rod that extends transversely of the shee-t material, the radiant energy that has been modulated by the shee-t material. Alternatively, a plurality of photocells may be mounted in a row exte~ding transversely of the web path for receiving radi~nt energy that has been modulated by the zone of the sheet section that corresponds with the field angle of a correspo~ding photocell.
~ According to one aspect of the invention there is provided in a method for in~pecting at least one longitudinal section of a moving sheet material for streaklike defectsl by repeatedly transversely scanning during a period P each such section of the material by means of radiant energy capable of being modulated by said sheet material, and by receiving said modulated energy on at least one photocell ~hereby to produce at said at least one photocell at least one train o measurement si~nals vm periodically recurring in response with the scanning `, ~ ' ' ~L~Z~ 9 of the material and which ~ary in intensity as an indication of the presence of said defects, wherein the S/N ratio of those having a S/N ratio of greater than one ~s improved by electronic correlation the improvemen-t wherein said corr~la-t:ion comprises -~he following steps:
delaying the measurement signals Vm in the train during one scan over a time delay t correspondi.ng with the scanning period P;
multiplying the measurement signals during a following scan with a factor that is proportional to the corresponding delayed signals, and is greater than l;
delaying the thus multiplied signals over said time delay t and multiplying the measurement signals produced during a subsequent scan with a fac-tor that is proportional to said corre-sponding previous multiplied signals, and so on for a number of scanning periods.
In accordance with a second aspect of the invention there is provided, in an apparatus for inspecting a moving sheet material for streaklike defects, comprising cooperating radiant energy means and photocell means arranged for periodically scanning said moving sheet material, means for amplifying the measurement signals Vm that are produced during each scan of the sheet material, and electronic correlation means for improving the S/N ratio of the measurement signals, the ;mprovement wherein said electronlc correlation means comprises:
delay lines controlled in response to the scanning of the material for delaying said measurement signals Vm over a delay _ _. ~ 7 _ t 7~
period equal to the scanning period P;
mult~pliers for multiply~ng the delayed siynals Vm with corresponding subsequent measurement signals Vm that occur during a following scanning period, and ~'or repeatlng such mulkiplica-tion for a number of periods -thereby to increase the S/N ra~io o~ those measurement signals having a signal component Vd greate~ than the effective noise component Vn, and means for evaluating said signals Vd with improved S/N
ratio thereby to identify streaklike defects in the sheet material.
Specific embodiments of the invention will now be described by way of example with reference to the accompanying drawings, in which:
Fig. 1 is a diagrammatic isometric view of an embodiment of a device for transversely scanning a moving web, Fig. 2 is a vertical section on line 2 2 of fig. 1, Fig. 3 is a section on line 3-3 of fig. 2, Fig. 4 is the electronic block circuit of -the device according to figs. 1 to 3, and Fig. 5 illustrates the sinusoidal displacement of the de-vice according to fig. 3, and the operative detection zone within said sinusoidal displacement, Fig. 6 is the basic block circuit of an additive correla-tor, and Fig. 7 is a diagram illustrating the S/N ratio improvement of the correlator of fig. 6, Fig. 8 is the basic block circuit of a multiplicative cor-relator, and Fig7 9 is a diagram illustrating the S/N ratio improvement of the correlator of fig. 8, - 7a -,, . ~
`
~ , , , ~ ~ Z~5~
~ig~ 10 is a more detailed block circuit of the multipli-ca-tive correlator of :Eig. 8, and ~ig. 11 is the electronic block circui-t of one em~odimen-t of a device according -to the present invention.
~ig. 12 is a representation o~ the screen of an oscillo-scope showing a noise sig~al, a de~ect signal and the combination of both signals, and ~ig. 13 is a representation on the same screen of the com-bination signal the S/~ ratio of which has been i~-proved by additive, respectively by multiplicative correlation.
~ igure 1 is a diagrammatic isometric view of a device for transversely scanning a moving web. ~he device com-prises a boxlike frame 20 that is slideably suspended by means of two guides 21 and 22 on two horizontally spaced, fixed parallel rods 23 and 24.
I-t will be understood that the illustrated suspension means may be replaced by any known other suspension or guide mechanism -that enables the frame to carry out a 20~ scanning movement in the direction indicated by the arrow ~ he transverse position of the frame with respect -to a web 10 that is advanced in the direction of the arrow 11 is controlled by a crank 26 that is pivotally fitted to one end of the frame and -to a drive wheel 27 that may con-tinuously ro-ta-te at a constant speed. ~he web 10 is passed through the device through a slotlike opening 28 along a path that is determined by the tangent plane to two rollers (not sho~m), one being located ups-tream and the other down-stream of the de~iee.
Within the device there are mounted several detectionunits next to each other at equal intervals, each covering a section of the width of the we~ One such uni-t is dia-grammatically illustrated in the vertical cross-sectional , ~V~1031 . .
..
.
: ~
~ , . .
~: ~. :
view of fig. 2, and the vertical longitudinal sectional view of ~igo 3.
~ he unit comprises two so-called fishtail op-tics 29 and 30 which are fiber op-tic assemblies wherein the fibers are bundled at one extremi-ty to a bund~e ~ith a circular cross-section, and ~herein the ~ibers diverge towards the o-ther extremity to form a line on which the single fibers are ranged closely adjacent to each other.
~he two fishtail assemblies have been ground at their lower ends to form an interface 31 a-t which they are adhered to each other. ~ rodlike lens 32 with a semi-circular cross-section projects the li~es of radiation at the lower ends of the fishtail optics on the web in the form of the two narrow elo~gate beams of radiant ener-gy hereinafter called "lines", that have been representedby the solid points 33 and 34, respectively. ~he radiant energy may be produced by any suitable source such as an incandescent bulb, a light-emitting diode (LE~), or the like. In the present case, two ~D's 95 and 96 were placed in face of the upper ends of the corresponding fishtail optics.
Radiation from both lines 33 and 34 tha-t is trans-mitted through the web may be bundled by a second rodlike lens, such as lens 35 onto an elongate photocell 36. rlhe sources of radiation as well as the photocell carry out an oscillating motion as indicated by the arrows 25.
The amplitude of said oscillation is that of the oscilla-ting amplitude of the frame 20 shown in fig. 3~
It should be noted that several units such as the one illustrated in fig. 2 are mounted at equal distances next to each other within the frame 20,and that the distan-ce between two successive units is smaller than the ampli-tude of oscillation of the frame in order to completel~
transversely scan the web. While -the web moves away from GV.1031 '~
- 10 _ the reader according to the illustration of fig. 2, -the web moves towards the right according to fig. 3, as indi-cated by the arrow 37.
'~he electronic block circuit of the device is shol~n in fig. 4. A web 10 tha-t moves in a direction normal ~o the plane of the drawing is scanned by pairs o~ ligh-t sources 95 and 96 that produce adjacent pairs of lines of radiant energy on the web as shown in detail in fig. 2.
Corresponding photocells 18 receive the radiation trans-mitted through the web and are bodily coupled as illustra-ted by the broken lines 40 and 41 with the ligh-t sources in order to carr~ out an oscillating displacement with respect to the web in the direc-tion of the arrow 25.
Both light sources of each pair of sources are alter-natively energized by a source 42 in response to an oscil-lator 43. ~he ou-tput signals Vp of the photocells are amplified by amplifiers L~ and synchronously demodulated by correspo~ding demodulators 45 that are likewise con-trolled by the oscillator 43. lhe amplifiers L~ may com- -prise a high-pass filter with a braakpoint at 20 Hz for 6 db attenuation.
~ he frame of the device is further provided with a pulse-generator L~6 in the form of a glass strip that is provided with a great number of spaced fields capable of periodicall~ absorbing the radia-tion from a stationary light source 47 that is directed to a photocell 48. ~he vertical dash and do-t line 49 represents a wall of the oscillating frame at which all electric terminals 50 may be grouped and connected by means of a flexible band com-prising different separate conductors with the remainingelectronic circuit of the device which in practice is mou~ted in a stationary column beside the oscillating frameO
It will be clear that upon oscillation of the frame 20 with the light sources and the photocells, no electric GV.1031 , :: . ... .
: .
. ~ .
~7;~5~
measurement signals Vm occur at the terminals 50 if a defect-free web is being examined. As a mat-ter of fac-t, the blocklike energizing of the pairs of light sources 95 and 96 produces outpu-~ signals oF the photocells ~Jhich have equal magnitude~ polarit~ and width so that s~Jnehro nous demodulation of them by -the circui-ts 45 produces a zero output voltage. In case however, a pair of light sources scan a streaklike defect, the a~plitude of the successive output signals of the corresponding photocell will no longer be equal for the two light sources during their passage over a line defect of the web so that the resulting demodulated signal will no longer be zero. ~he same is -true for the other pairs of light sources and the corresponding photocells -that each cover a section of the width of the web, as indicated by the distance c.
In connection with this distance c, it should be noted that this is a fraction only of the peak to peak amplitude of the oscillation of the frame 20. This is illustrated in the diagram of figo 5 wherein -the distance 20~ d represents the peak to peak amplitude of the frame 20, whereas the distance c represents the effective scanning width of each photocell. As a consequence of the mentioned difference between d and c, the distance c covers a portion of the curve 51 -that represents the sinusoidal motion of the frame, that deviates not quite much from a line, where-by the corresponding speed of the frame during measurin~
is nearly uniform.
As a consequence of the frame amplitude d that is greater than the distance c between two adjacent measuring ~0 units, overlapping measurement occurs. ~he electronic circuit is, however, so arranged (not shown) that the ou-t-put signals of overlapping measurements of adjacent web sections are not processed.
GV.1031 ' ..:' - 12 ~
~ he right hand portion of the drawing of fig. 4 shol,Js a known additive correlation circuit for improving the S/~ ratio o~ the measurement signals ~m. I-t should be noted -that actually there is one such correlation ci,rGuit for each pho-tocelL circuit.
~ he electronic correlation circuit comprises an adder-type amplifier 521 an analogue to digi-tal conver-ter 53, a delay line 54 in -the form of an 8 bit shift register wi-th 1024 positions, a digital to analogue converter 55, and feedback element 56. ~he adder sums the incoming signals Vi with a delayed output signal ~0 which has ~een multi-plied by a feedback gain which is strictly less -than uni-ty.
~he delay line 54 is controlled in synchronism with the scanning of the web by pulses produced by the pulse gene-rating strip 46 that comprises in this specific embodiment1024 pairs of light absorbing and light transmission areas.
The signals of the photocell 48 are appropriately shaped by a pulse shaper 57.
~he opera-tion of the circuit is as follows. An input signal that occurs at a time t1~ see fig. 5, is stored in the first stage of the shift register 54 until at a time t2 the pulse generator 46 controls the shift register to shift its informa-tion over one place whereby the first signal is transferred to a second position, and the new signal that occurs at the time t2 is entered into the shift register. ~he same procedure continues un-til the generator 46 has been displaced over a distance c i.e. at -the time t1024 which means that the scanning of the respective sec-tion of the web has been comple-ted.
'~he circuit is so arranged (not illus-tra-ted) that during the returning movement of the frame the shift registers remain inoperative. At the second scanning of the web, a pulse from the generator 46 at a time t1 transfers the signal of the shift register that is at the position 1024 GV.1031 " .
:
to the posi-tion 1 whereby (through 55, 56, 52 and 53) said signal is added -to the next signal that occurs a-t the same transverse zone of the web as -the zone where said t-ransferred signal occurred~ etc. The same operation is repeated a number of times thereby -to increa~e the S/~
ra-tio of each of the signals a-t the posi-tlons ~rom 1 -to 1024 by a factor as will further be explained4 '~he ou-tput signals ~0 of each correlator circuit are fed to a trigger 58 where they are compared with a re~e-rence signal from a reference signal generator 59. Incase the difference between both signals exceeds a pre~
determined value, -the output signal at terminal 63 may, either directly control a warning mechanism, or it may be stored either in analogue or in converted digital form in a memory together with the exact address of the location of the defect on -the web. One part of said address is formed by the pulse number of the generator 46 which indi-cates the -transverse location of -the defect on the web.
Another part of the address may be formed by the number from a counter that counts the number of meters of the web which have been unwound from a roll. Such counter is usually reset at zero at the starting of the treatmen-t, for example coating, of each new web roll. S-till another part of the address may be formed by an identification number for the device that detected the object~ ~he cir-cuit comprises further a second trigger 61 tha-t is direc-t-ly connec-ted with the inpu-t terminal 50 and -that serves for the instant signalling at a -terminal 62 of defects that are so severe that they need not a repeated circula-tion through the correlation circuit~ with the correspon-ding delay, to reach a level at which s~fficient signal to noise distinction has been obtained.
It should finally be noted tha-t the active operation of the device, namely from a time t1 to a time t1024, is a fraction only of the period of one complete scanning~
GVo 1 031 . -.
.
, ~Z'7%5~
1~
The examination of the moving web occurs in consequence interruptedlyconsidered in the longitudinal direction. Taking into account however, that the device according to the invention is intended for the identification of streaklike defects of protracted duration, this lack of continuity of the measurement in the longitudinal direction of the web is without any importance.
Further details about the principle of operation of the device described hereinbefore, namely the scanning of a material across at least part of the width of its path by means oF one or more beams of lo radiation which or each of which at any given instant irradiates a narrow elongate zone (elongate in the d;rection of travel of the material) within the width of the said path, receiving quanta of such radiant energy, modulated by adjacent successively irradiated zones of the material, on a photocell, deriving from such photocell a multiplicity of sequential signals each representative of a particular said zone, and comparing successive said signals with each other and using amplitude differences between compared signals as indicative of the presence of said defects, may be found in our Canadian application no. 330,537 filed June 26, 1979 by Agfa-Gevaert N.V.
The basic operation of an additive correlator as used in the block circuit of fig. 4 is illustrated in figs. 6 and 7.
Referring to fig~ 6 the circuit comprises an adder 679 a delay line 68 with a delay time 2~ and an adjustable feedback element 69 with a gain A comp~ised between G and 1~ It can be shown that the relation between the output voltage V0 and the input voltage V
is expressed by :
1 l-A t 30VO ~ Vj 1~ e~ ~ ) (1) GV.1031 ~ ~ , : '
A ~--V. L~
~ his invention relates to a method and a device for inspecting a moving sheet material for streaklike defects, comprising means for directing radiant energy on one sur-i face of the sheet material and photocell means for measu-ring at distinct lateral positions of the sheet material ! the intensity of the radia-tion after transmission of -the - radiation through, or after reflection on the surface of the sheet material.
Enown devices operate satisfactorily for the detection of rather ooarse speck- and streaklike defects in moving sheets since in such cases the signal-to-noise ratio of the output sig~als of the photocells is sufficien-tly great to enable the signals, after their amplification, to be com-pared with a predetermined reference signal in order to identify a defect. Suchlike defect identifying sig~als ma~ either be stored in digital or analogue form in a memory together with the exact address of the defect, or they may directly be signalled as in the coating of webs in order to arrest or to re-adjust the coating, or they ma~ ~directly control a sorting mechanism as in the inspec-tion of sheets cut from a coated web. As examples of rather coarse defects can be mentioned specklike defects caused by -the inclusion o~ a stra~ge object, e~g. a dust particle in a coated layer, or a dry spot, e~g~ an air GV~10~1 ~.
- -', .~ .: ' ~_~ 2~f~
-- 2 --bubble included duri~g the coa-ti~g of a layer on a support.
In those cases, however7 where very minute defects have to be identified, fluttering of the material at the place of measurement a~d ~oise of the electronic circui-t~ may cause variatio~s in the uutput signals o~ -the photocells that are of the same order of magnitudo as those caused by a sheet defect.
I-t is known to use electronic correlatio~ techniques for improving the signal-to-~oise ratio, hereinafter re-ferred to as S/~ ratio, of the defect identifying signals.~ basic condition for the operation of such techniques is that the defect to be identified should be recur~ive during a number of, say at least 5, successive scans.
~his means that the defects have to ke of protracted dura-tion in the machine direction, that îs in the coating direc-tio~ in the case of the inspection of coated la~ers. In other words, only streaklike defect signals are suitable for ~/N ratio improvement.
The use of additive correlation in apparatus for the inspection of movi~g webs is disclosed for instance in G~-P
1,471,316 filed June 4, 1974 b~ Eastman Eod~k Co. and in U~P
~005,281 of Mark ~dwi Faulhaber a~d Edmund Haislett Smith Jr.
issued Januar~ 25, 19770 T~e mentioned use does not provide results that are e~tirely satisfactory for the following reasons~
~he ~/N ratio improvement that can be obtained is rather limited as will be further explai~ed in detail in the description of the invention. ~or instance, for a gain in the feedback loop of a delay line equal to 0.907 the S/~ gain is 12.78 db whereas for A = 0.99, the gain is 22~98 d~ in accordance with the formula : S/~ (in db) =
~0 log ~
Further, ~hereas it is possible to use for a factor = 0.90 an analogue circuitry7 the use of a factor A =
0~99 inevitabl~ ~ecessitates the use of a digital circuitry ~. Z~7~25~ .
in order to obtain a sufficient accuracy, and suchlike digi-tal arrangement requires a rather ex-tensive hardware layout.
Finally, the time t o:E the correlation clrcuit f'or reaching a significan-t S/~ ratio improvement is rathex important so tha-t in some cases i-t may ta'ke up -to ~00 seconds and more before -the correlated signal is availa-ble. It is the objec-t of the invention -to provide a new correlation technique in the inspection of a moving sh0et material for streaklike defects, that does not show the shortcomings menticned hereinbefore. ~he invention aims in particular at eonsiderably improving the S/~ ratio of photocell signals that are encountered in the identifica-tion of very subtle streaklike defects in moving sheet materials.
As an example of very subtle streaklike defects~
streaklike irregularities may be mentioned that occur at 'the coating on a web of a layer,,or of a combination of layers as the case may be, by means of a so~called cascade 20~ or'slide hopper coater. In certain applieation fields~
notably in photography, very stringent requirements are put to the uniformity of a coated layer, and it is shown that defect signals resulting from streaklike thickness 'deviations of an order of magnitude under 0.1 %, even when 25, additively correlated, are not discernable from the noise eomponent of the signals.
' ' As to the oceurrence of the mentioned defect in the said coating teehnigue, it is assumed that the most impor-,tant~eause for such defect is the passage of a web splice through the coa-ting bead. It appears that a web splice , entrains,air in the coating bead and that such air may re-main there for quite long periods, in some circums-tances eovering hundreds of meters of material, so that the coa-t-ing thickness at tha-t area is correspondingly dis-turbed.
G~.1031 :
~-~z~
_ 4 Considering the fact that in the manufacture of delicate materials, such as for instance radiographic film or film for graphic reproduction techniques, thickness deviations in coated layer in the order of magnitude of 0.1 % and preferably even less than 0.1 %
should be indentified, it will be understood that the signal--to-nolse characteristics of common detection systems do not enable a rel~able and rapid retrieval of the mentioned defect.
In accordance with the present invention~ a method for ;nspect;ng at least one longitudinal section of a moving sheet material for lo streaklike defects, by repeatedly transversely scanning said at least one section of the material by means of radiant energy capable of being modulated by said sheet material, and by receiving said modulated energy on at least one photocell thereby to detect at said at least one photocell signals that are periodically recurring in response with the scanning of the material and the variation of which signals is indicative of the presence of said defects, comprises improving the S/N ratio of those said periodically recurring signals the S/N ratio of which is greater than one, by delaying each measurement signal Vm over a time delay t corresponding with the scanning period P, multiplying a next signal Vm with a factor that is proportional to the product of said delayed signal and a factor A, delaying said multiplied signal over said time delay t and multiplying the next signal Vm with a factor that is proportional to the product of said previous multiplied signal and said factor A~
and so on for a number of times, wherein said factor A is at least equal to one, and wherein further Vm.A ~ 13 for said signal Vm being expressed in volts.
The notion "streaklike defect in a moving sheet material" stands in the present statement for a line-wise defect that runs parallel with the direction of movement of the sheet material during its inspection.
GV.1031 .{:r ~
-The term "sheet material" stands for webs that may have a length up to many hundreds of meters~ as well as for a plurality of distinct sheets that may have been cut from such webs and -that may be inspec-ted in succe~sion.
The term "modulated by" indicates that radia-~ion is attenuated on its pa-th towards the photocell, either b~
transmission through the sheet material or by reflectior from the surface of said sheet material~ ~aid defects modify -the attenuation.
~he expression "signals wi-th a ~/~ ratio greater than one" means in the present specification signals the defect-component of which is greater than the effective noise component. ~his means that peak-to-peak noise levels may be presen-t in the photocell signal that are much greater than the useful signal level but the effective value of which, that is the rms value determined over a time that is equal to the quo-tient of -the scanning time t by the number of discrete measurements during said time~
~he scanning period "P" is the time comprised between two successi~e scans of a given zone of said at leas-t one longitudinal section from edge to edge in one direction.
~ he measurement signal Vm may be the output signal of a photocell that must be compared with a given reference signal in order to establish the occasional deviation of said measurement signal, thereby to detect a defect at the lateral position of the sheet material inspec-ted by the photocell, but said signal Vm may also be~ and in a different embodiment preferably is, the difference, amplified as the case may be, between two successive signals of a mûltiplicity of sequential output signals that re~
present the scanning of the sheet material.
~ he term "scanning" stands for the repea-ted inspection of the ~heet material in a direction that runs -transverse with respect to the direction of movement of the sheet ma G~.10~1 ~2,7~5~
terial. ~he scanning may occur by means o~ a beam of ra-diatio~ that is projected on the sheet ma-terial and that is swept over the material. ~his sweeping may be done by means of a source of radiation mounted on a frame arranged for oscillation transverse wi-th regpect to -the m~-teIial, but said sweeping may also be carried out by mearls o~' a beam of radiation that is de~lec-ted by a rotati~g or pivo-tir~g mirror or -the like. Alterna-tively, ~he scannin~ may occur b~ means o~ a plurality o~ dis-tinct small light sources, for insta~ce light emitting diodes (~D's) that are mounted closely adjacent to each other in a row and that may be energized in sequence to produce a radia-tion spot that scans the sheet material.
~ he photocell may be arra~ged for lateral displacement sy~chronous with the scanning beam of radiation, but the photocell may also collect, for instance by means of a collecting glass rod that extends transversely of the shee-t material, the radiant energy that has been modulated by the shee-t material. Alternatively, a plurality of photocells may be mounted in a row exte~ding transversely of the web path for receiving radi~nt energy that has been modulated by the zone of the sheet section that corresponds with the field angle of a correspo~ding photocell.
~ According to one aspect of the invention there is provided in a method for in~pecting at least one longitudinal section of a moving sheet material for streaklike defectsl by repeatedly transversely scanning during a period P each such section of the material by means of radiant energy capable of being modulated by said sheet material, and by receiving said modulated energy on at least one photocell ~hereby to produce at said at least one photocell at least one train o measurement si~nals vm periodically recurring in response with the scanning `, ~ ' ' ~L~Z~ 9 of the material and which ~ary in intensity as an indication of the presence of said defects, wherein the S/N ratio of those having a S/N ratio of greater than one ~s improved by electronic correlation the improvemen-t wherein said corr~la-t:ion comprises -~he following steps:
delaying the measurement signals Vm in the train during one scan over a time delay t correspondi.ng with the scanning period P;
multiplying the measurement signals during a following scan with a factor that is proportional to the corresponding delayed signals, and is greater than l;
delaying the thus multiplied signals over said time delay t and multiplying the measurement signals produced during a subsequent scan with a fac-tor that is proportional to said corre-sponding previous multiplied signals, and so on for a number of scanning periods.
In accordance with a second aspect of the invention there is provided, in an apparatus for inspecting a moving sheet material for streaklike defects, comprising cooperating radiant energy means and photocell means arranged for periodically scanning said moving sheet material, means for amplifying the measurement signals Vm that are produced during each scan of the sheet material, and electronic correlation means for improving the S/N ratio of the measurement signals, the ;mprovement wherein said electronlc correlation means comprises:
delay lines controlled in response to the scanning of the material for delaying said measurement signals Vm over a delay _ _. ~ 7 _ t 7~
period equal to the scanning period P;
mult~pliers for multiply~ng the delayed siynals Vm with corresponding subsequent measurement signals Vm that occur during a following scanning period, and ~'or repeatlng such mulkiplica-tion for a number of periods -thereby to increase the S/N ra~io o~ those measurement signals having a signal component Vd greate~ than the effective noise component Vn, and means for evaluating said signals Vd with improved S/N
ratio thereby to identify streaklike defects in the sheet material.
Specific embodiments of the invention will now be described by way of example with reference to the accompanying drawings, in which:
Fig. 1 is a diagrammatic isometric view of an embodiment of a device for transversely scanning a moving web, Fig. 2 is a vertical section on line 2 2 of fig. 1, Fig. 3 is a section on line 3-3 of fig. 2, Fig. 4 is the electronic block circuit of -the device according to figs. 1 to 3, and Fig. 5 illustrates the sinusoidal displacement of the de-vice according to fig. 3, and the operative detection zone within said sinusoidal displacement, Fig. 6 is the basic block circuit of an additive correla-tor, and Fig. 7 is a diagram illustrating the S/N ratio improvement of the correlator of fig. 6, Fig. 8 is the basic block circuit of a multiplicative cor-relator, and Fig7 9 is a diagram illustrating the S/N ratio improvement of the correlator of fig. 8, - 7a -,, . ~
`
~ , , , ~ ~ Z~5~
~ig~ 10 is a more detailed block circuit of the multipli-ca-tive correlator of :Eig. 8, and ~ig. 11 is the electronic block circui-t of one em~odimen-t of a device according -to the present invention.
~ig. 12 is a representation o~ the screen of an oscillo-scope showing a noise sig~al, a de~ect signal and the combination of both signals, and ~ig. 13 is a representation on the same screen of the com-bination signal the S/~ ratio of which has been i~-proved by additive, respectively by multiplicative correlation.
~ igure 1 is a diagrammatic isometric view of a device for transversely scanning a moving web. ~he device com-prises a boxlike frame 20 that is slideably suspended by means of two guides 21 and 22 on two horizontally spaced, fixed parallel rods 23 and 24.
I-t will be understood that the illustrated suspension means may be replaced by any known other suspension or guide mechanism -that enables the frame to carry out a 20~ scanning movement in the direction indicated by the arrow ~ he transverse position of the frame with respect -to a web 10 that is advanced in the direction of the arrow 11 is controlled by a crank 26 that is pivotally fitted to one end of the frame and -to a drive wheel 27 that may con-tinuously ro-ta-te at a constant speed. ~he web 10 is passed through the device through a slotlike opening 28 along a path that is determined by the tangent plane to two rollers (not sho~m), one being located ups-tream and the other down-stream of the de~iee.
Within the device there are mounted several detectionunits next to each other at equal intervals, each covering a section of the width of the we~ One such uni-t is dia-grammatically illustrated in the vertical cross-sectional , ~V~1031 . .
..
.
: ~
~ , . .
~: ~. :
view of fig. 2, and the vertical longitudinal sectional view of ~igo 3.
~ he unit comprises two so-called fishtail op-tics 29 and 30 which are fiber op-tic assemblies wherein the fibers are bundled at one extremi-ty to a bund~e ~ith a circular cross-section, and ~herein the ~ibers diverge towards the o-ther extremity to form a line on which the single fibers are ranged closely adjacent to each other.
~he two fishtail assemblies have been ground at their lower ends to form an interface 31 a-t which they are adhered to each other. ~ rodlike lens 32 with a semi-circular cross-section projects the li~es of radiation at the lower ends of the fishtail optics on the web in the form of the two narrow elo~gate beams of radiant ener-gy hereinafter called "lines", that have been representedby the solid points 33 and 34, respectively. ~he radiant energy may be produced by any suitable source such as an incandescent bulb, a light-emitting diode (LE~), or the like. In the present case, two ~D's 95 and 96 were placed in face of the upper ends of the corresponding fishtail optics.
Radiation from both lines 33 and 34 tha-t is trans-mitted through the web may be bundled by a second rodlike lens, such as lens 35 onto an elongate photocell 36. rlhe sources of radiation as well as the photocell carry out an oscillating motion as indicated by the arrows 25.
The amplitude of said oscillation is that of the oscilla-ting amplitude of the frame 20 shown in fig. 3~
It should be noted that several units such as the one illustrated in fig. 2 are mounted at equal distances next to each other within the frame 20,and that the distan-ce between two successive units is smaller than the ampli-tude of oscillation of the frame in order to completel~
transversely scan the web. While -the web moves away from GV.1031 '~
- 10 _ the reader according to the illustration of fig. 2, -the web moves towards the right according to fig. 3, as indi-cated by the arrow 37.
'~he electronic block circuit of the device is shol~n in fig. 4. A web 10 tha-t moves in a direction normal ~o the plane of the drawing is scanned by pairs o~ ligh-t sources 95 and 96 that produce adjacent pairs of lines of radiant energy on the web as shown in detail in fig. 2.
Corresponding photocells 18 receive the radiation trans-mitted through the web and are bodily coupled as illustra-ted by the broken lines 40 and 41 with the ligh-t sources in order to carr~ out an oscillating displacement with respect to the web in the direc-tion of the arrow 25.
Both light sources of each pair of sources are alter-natively energized by a source 42 in response to an oscil-lator 43. ~he ou-tput signals Vp of the photocells are amplified by amplifiers L~ and synchronously demodulated by correspo~ding demodulators 45 that are likewise con-trolled by the oscillator 43. lhe amplifiers L~ may com- -prise a high-pass filter with a braakpoint at 20 Hz for 6 db attenuation.
~ he frame of the device is further provided with a pulse-generator L~6 in the form of a glass strip that is provided with a great number of spaced fields capable of periodicall~ absorbing the radia-tion from a stationary light source 47 that is directed to a photocell 48. ~he vertical dash and do-t line 49 represents a wall of the oscillating frame at which all electric terminals 50 may be grouped and connected by means of a flexible band com-prising different separate conductors with the remainingelectronic circuit of the device which in practice is mou~ted in a stationary column beside the oscillating frameO
It will be clear that upon oscillation of the frame 20 with the light sources and the photocells, no electric GV.1031 , :: . ... .
: .
. ~ .
~7;~5~
measurement signals Vm occur at the terminals 50 if a defect-free web is being examined. As a mat-ter of fac-t, the blocklike energizing of the pairs of light sources 95 and 96 produces outpu-~ signals oF the photocells ~Jhich have equal magnitude~ polarit~ and width so that s~Jnehro nous demodulation of them by -the circui-ts 45 produces a zero output voltage. In case however, a pair of light sources scan a streaklike defect, the a~plitude of the successive output signals of the corresponding photocell will no longer be equal for the two light sources during their passage over a line defect of the web so that the resulting demodulated signal will no longer be zero. ~he same is -true for the other pairs of light sources and the corresponding photocells -that each cover a section of the width of the web, as indicated by the distance c.
In connection with this distance c, it should be noted that this is a fraction only of the peak to peak amplitude of the oscillation of the frame 20. This is illustrated in the diagram of figo 5 wherein -the distance 20~ d represents the peak to peak amplitude of the frame 20, whereas the distance c represents the effective scanning width of each photocell. As a consequence of the mentioned difference between d and c, the distance c covers a portion of the curve 51 -that represents the sinusoidal motion of the frame, that deviates not quite much from a line, where-by the corresponding speed of the frame during measurin~
is nearly uniform.
As a consequence of the frame amplitude d that is greater than the distance c between two adjacent measuring ~0 units, overlapping measurement occurs. ~he electronic circuit is, however, so arranged (not shown) that the ou-t-put signals of overlapping measurements of adjacent web sections are not processed.
GV.1031 ' ..:' - 12 ~
~ he right hand portion of the drawing of fig. 4 shol,Js a known additive correlation circuit for improving the S/~ ratio o~ the measurement signals ~m. I-t should be noted -that actually there is one such correlation ci,rGuit for each pho-tocelL circuit.
~ he electronic correlation circuit comprises an adder-type amplifier 521 an analogue to digi-tal conver-ter 53, a delay line 54 in -the form of an 8 bit shift register wi-th 1024 positions, a digital to analogue converter 55, and feedback element 56. ~he adder sums the incoming signals Vi with a delayed output signal ~0 which has ~een multi-plied by a feedback gain which is strictly less -than uni-ty.
~he delay line 54 is controlled in synchronism with the scanning of the web by pulses produced by the pulse gene-rating strip 46 that comprises in this specific embodiment1024 pairs of light absorbing and light transmission areas.
The signals of the photocell 48 are appropriately shaped by a pulse shaper 57.
~he opera-tion of the circuit is as follows. An input signal that occurs at a time t1~ see fig. 5, is stored in the first stage of the shift register 54 until at a time t2 the pulse generator 46 controls the shift register to shift its informa-tion over one place whereby the first signal is transferred to a second position, and the new signal that occurs at the time t2 is entered into the shift register. ~he same procedure continues un-til the generator 46 has been displaced over a distance c i.e. at -the time t1024 which means that the scanning of the respective sec-tion of the web has been comple-ted.
'~he circuit is so arranged (not illus-tra-ted) that during the returning movement of the frame the shift registers remain inoperative. At the second scanning of the web, a pulse from the generator 46 at a time t1 transfers the signal of the shift register that is at the position 1024 GV.1031 " .
:
to the posi-tion 1 whereby (through 55, 56, 52 and 53) said signal is added -to the next signal that occurs a-t the same transverse zone of the web as -the zone where said t-ransferred signal occurred~ etc. The same operation is repeated a number of times thereby -to increa~e the S/~
ra-tio of each of the signals a-t the posi-tlons ~rom 1 -to 1024 by a factor as will further be explained4 '~he ou-tput signals ~0 of each correlator circuit are fed to a trigger 58 where they are compared with a re~e-rence signal from a reference signal generator 59. Incase the difference between both signals exceeds a pre~
determined value, -the output signal at terminal 63 may, either directly control a warning mechanism, or it may be stored either in analogue or in converted digital form in a memory together with the exact address of the location of the defect on -the web. One part of said address is formed by the pulse number of the generator 46 which indi-cates the -transverse location of -the defect on the web.
Another part of the address may be formed by the number from a counter that counts the number of meters of the web which have been unwound from a roll. Such counter is usually reset at zero at the starting of the treatmen-t, for example coating, of each new web roll. S-till another part of the address may be formed by an identification number for the device that detected the object~ ~he cir-cuit comprises further a second trigger 61 tha-t is direc-t-ly connec-ted with the inpu-t terminal 50 and -that serves for the instant signalling at a -terminal 62 of defects that are so severe that they need not a repeated circula-tion through the correlation circuit~ with the correspon-ding delay, to reach a level at which s~fficient signal to noise distinction has been obtained.
It should finally be noted tha-t the active operation of the device, namely from a time t1 to a time t1024, is a fraction only of the period of one complete scanning~
GVo 1 031 . -.
.
, ~Z'7%5~
1~
The examination of the moving web occurs in consequence interruptedlyconsidered in the longitudinal direction. Taking into account however, that the device according to the invention is intended for the identification of streaklike defects of protracted duration, this lack of continuity of the measurement in the longitudinal direction of the web is without any importance.
Further details about the principle of operation of the device described hereinbefore, namely the scanning of a material across at least part of the width of its path by means oF one or more beams of lo radiation which or each of which at any given instant irradiates a narrow elongate zone (elongate in the d;rection of travel of the material) within the width of the said path, receiving quanta of such radiant energy, modulated by adjacent successively irradiated zones of the material, on a photocell, deriving from such photocell a multiplicity of sequential signals each representative of a particular said zone, and comparing successive said signals with each other and using amplitude differences between compared signals as indicative of the presence of said defects, may be found in our Canadian application no. 330,537 filed June 26, 1979 by Agfa-Gevaert N.V.
The basic operation of an additive correlator as used in the block circuit of fig. 4 is illustrated in figs. 6 and 7.
Referring to fig~ 6 the circuit comprises an adder 679 a delay line 68 with a delay time 2~ and an adjustable feedback element 69 with a gain A comp~ised between G and 1~ It can be shown that the relation between the output voltage V0 and the input voltage V
is expressed by :
1 l-A t 30VO ~ Vj 1~ e~ ~ ) (1) GV.1031 ~ ~ , : '
2~
and that the improvement of the S~N ratio is :
~/N - ~ or, expressed in db : (2) ~/N = 10 log 1 A ~3) ~ he s~stem reaches 99 ~ of its maximum ~/N improve~
men-t after a -time t given by :
t = ~ (4~
As mentioned already in -the in-troduc-tio~ of the speci-fication, a significant improvement of -the S/N ra-tio can only be ob-tained for a gain close to 1, e~g. 0.99, what requires in practice a complete digital lay-out of the system in order to obtain a-stable operation. The time re~
quired for reaching such improvement is not neglectable.
~or instance, for the device illustrated in figs~ 1 to 4 wherein ~ is equal to 0.2 s,t = 100 (from t = ~ 9-).
~ ig. 7 is a diagram illustrating the S/~ improvement for A = O.99,~ = 0.2 s, and t = 100 s. ~he S~N improve-ment amounts to a ra-tio of 14.106. ~he abscissa is a time axis in seconds whereas the ordinate represents V0. The maximum level of the noise component Vn of -the output signal V after 100 s has been arbitrarily illustra-ted at a o ~
level of 1 volt. In fact Vn = Vi~ ~ whereas d l ~ . It may be seen that both the noise signal Vn and the defect signal Vd exponentially increase in the same direction towards a determined limit.
In accordance with the present invention, -the ~/N
ratio of periodically recurring signals is improved by multiplicatively correlating said signals instead of additively correlating them~ ~he basic opera-tion of multi-plic~tive correlation is illustrated in figo 8 wherein the block circuit comprises a multiplicator 70, a delay line 71 with a delay time ~ , and a feedback element 72 with a gain factor A~ It can be shown that the relation between the output voltage ~0 and -the input voltage Vi is expressed GV.1031 :,.
.
~z~
by: [VO = i_O ~ Vi (t - i.t) ~ (5) In consequence, -the sys-tem produces a S~N improve~ent that is infinitely grea-t in -theory. In practice~, said improvemen-t is limited by the maximal level -the signal component Vd can reach in the circuit. ~ommon values for said signal Vd are within 12 and 15 volts in modern inte-grated circuit lay~outs.
A basic condition for the operation of -the described circuit is -that the product of -the input signal Vi with its delayed equivalent is always greater than oneO In case said product is smaller than one, then repea-ted multipli-cation will only reduce the amplitude of the signal instead of increasing it. ~he mentioned condition may be met in one of two ways : either the signal Vi is greater than 1 volt and in that case the gain A may be equal to one, or -the signal Vi is smaller than one volt and in that case the gain A is adjusted so that the produc~t Vi.A is greater than 1.
Another basic condition for the operation of the cir-cuit is that the ratio of the useful signal level Vd to,-the effective noise level Vn of the input voltage Vi be greater than one. Otherwise the system is incapable of'making a dis-tinction between signal and noise. ~he operation of the system is illustrated in the diagram of fig. 9 wherein the abscissa is a time axis in seconds, and the ordinate ,represents the output voltage VO. In case Vi = 1 volt and for A = 1, there is no S/N improvement whatsoever.
In case Vi~ 1, -than it may be seen that as a conse-quence of repeated multiplication the outpu-t voltage VO
increases towards infinity. In practice said vol-tage VO
will reach a limit Vsat which is the saturation vol-tage or the full scale output voltage of the electronic circuitO
In case Vic 1, then the output voltage decreases as a con-sequence of repeated multiplication and tends towards zero.
GV.1031 ', ?,7Z~
Thus it is clear that the signal component Vd of Vi should be greater than 1 whereas the effective noise componen-t Vn f Vi should be smaller -than 1.
~ig. 10 is a block circuit of a mul-tiplicative corre-lator wherein common integrated circuits are u~ed~ ~ mul~i-plier 75 (type 801~ manufactured by Intersil Corp.) is followed by an analogue to digi-tal convertor 76 (type AD 7570 manufactured by ~nalog Devices), a digital 8-bits shift register 77 (-type SY 2533 manufactured by Synertek) ,k,:,",, 10 and a digital to analogue convertor ~ (type AD 559 manufac-tured by Analog Devices). ~he circuit comprises further an analogue adder 79 (type ~A 741 manufactured by ~airchild) for adding to the ou-tput voltage V0 a small voltage Va that may be set by means of a potentiometer 80.
~he purpose of said voltage Va is as follows. At the starting of a correlation cycle, it may occur that, depend-ing on the offset characteristics of the convertor 78, that the output voltage V0 is zero. ~he repeated multipli-cation of V0 by Vi will involve no result since multiplica-20~ tion by a factor zero gives a product that remains zeroO
~ herefore, in order to take account of said possibility that the outpu-t voltage V0 might be zero, a given voltage Va is added to the voltage V0 before the multiplication with Vi occursO Said voltaga Va may be small since its only purpose is -to make sure that Vi will not become multiplied by zero. I~ practice, Va will be chosen between 100 and 200 mV for a full scale output voltage of the device of 10 V. It is clear that the addition of said voltage Va "falsifies" so to say the correlated output signal V0.
It should be understood, however, that this "falsification"
is neglectable as compared by the dis-tortion of the signal Vi as a consequence of the multiplicative correlation. As a matter of fact, any signal Vi that comprises a signal component greater than 1 volt and a noise component smaller GV.1031 :
;' `,j :
ZJ~ ~ 5 than 1 volt, inevitably is amplified after a number of correla-tions to the saturation voltage Vsat of -the device.
Thus, whether the signal componen-t of Vi is, for instance, 1.1 or 1.5 V, the output vol-tage V0 will become within -the shortest time equal to Vsat. r~he described dis~or-~ion of the input signal Vi i5 comple-tely imma-terial in this type of apparatus wherein the o~ly aim is -to increase -the S/~ ratio of' a signal Vi that fulfils the basic conaitions set forth hereinbefore, as rapidly and as much as possible at all.
In case the signal component Vd of a signal Vi is smaller than 1 V, it is yet possible to multiplicatively correlate said signal in amplifyi~g it by a factor A so ~-that the product of Vi.A is greater than 1. ~n amplifier for carryi~g out the mentioned s-tep is illustrated in bro-ken lines by the block 81 in fig~ 10. ~aid block might as well figure between the input terminal 82 of the circuit and the multiplier 75, between the multiplier 75 and the convertor 76, etc., thereby to provide at the lower input 20~ terminal of the multiplier 75 an input signal greater than 1 V.
~ ig. 11 illustrates the device described hereinbefore with reference -to figs. 1 to 4, modified`however~ for mul-tiplicative correlation of the measureme~t signal Vm. ~he part of the drawing situated at the lef-t hand side of the dash and dot line 49 is identical with that of fig. 4 and has therefore been given the same re~erence numerals, except for the amplifiers 44 the gain o~ which has been made adjustable, as indicated by the conventional sign, so that the output measurement signals Vm at terminals 50 may be so adjusted that the noise of the signal at each terminal 50 is just smaller than 1 volt.
The signals Vm form in fact the input signals Vi o~
the multiplicative correlator, the shift register 77 of GVo 1031 .
~ . .
. ~ .
31 ~27 19 ~-which is controlled synchronously with the scanning of the different sections of the web by means of the ou-tput sig~al from the pulse shaper 57. ~he correlator is followed by two comparators 58 and 61 -that have the same function as the comparators disclosed in fig. 4. ~le source 80 has been adjusted for adding a signal Va = 0~2 V
to the output sig~al VO. ~he following da-ta illustrate the operation of the described device that was used for the inspection of a wet coating of silver halide that was applied by means of a slide hopper to a polyethylene terephthalate film having a width of 1770 m.
Scanning amplitude d of the frame 10 : 35 cm ~ffective scanning width c : 21.5 cm -Oscillation period P : 2.4 s 15 ~ffective scanning time ~ : 0.6 s ~umber of optical units : 8 Width of a line of radiant energy on the web : 1 mm ~ength of a line of radiant energ~ : 10 cm 20 Distance between the two lines of a pair of lines : 1 mm Frequency (f) of -the oscillator 43 (that is also the inspection frequency) : 10 kHz Number of signal comparisons during one scan (~xf) : 6000 2 ~umber of signal observations during 5 one scan : 1024 ~umber of signal circulations through a correlator : 5 Corresponding observation time : 12 s S/W improvement obtained by additive correlation according to fig. 4 : 12.79 db (for attenutation A = 0.9) S/N improvement obtained by multiplica-tive correlation according to fig~ 11 0 42.14 db (for 8 bit resolution of the AD
convertor) GV.1031 . .
: ~ ' ,'7 Variation in the output signals for a film fluttering of 1 mm at a frequency above 20 Hz : 0.5 %
Components used :
~amps 35 and 36 : ~D type G~ manufactured by ~'IRE DIVIS:CON
Inter~ational Audio Visual Inc.
Photocells 18 : PIN-LL~ manufactured by Unl-ted De-tector rL~ech-nology Spectral band of measurement : 880 nm wavelength.
It should be noted that the ~ul-tiplici-ty of signal comparisons, namely 6000, -that is carried out by the measuring device with the alternately operated pairs of lamps 95 and 96, is much greater than the number, namel~
1024, of discrete signals into which the output voltage of the measuring device is divided for carrying out the correlation. It can thus be said that finally the re-solving power of the device is 102~ points for each websection being examined. Whereas this number is largely sufficient for many applications, it should be understood it can easily be increased by using instead of pulse strip ~6 another pulse strip capable of producing a larger num-ber of pulses. It is evident that the number of the posi-tions of the shift register 77 should correspondingly be increased. Another point that has to do with the sensiti-vit~ of the device is that -the ~DC convertors 75 and the DAC convertors 78 are 8 bit devices whereby 256 (that is 28) distinct levels of the signals ~i can be transmi-tted~
It is clear that by using convertors with a number of bits 2n, wherein n is greater than 8, the sensitivity for small signal variations can further be increased.
In order to visualize the ~/N ratio improvement that can be ob-tained in accordance with the present invention, the following test was made. A generator for generating a white noise containing frequencies within the range of O to 20 kffz was used to similate the input voltage ~i f a correlator. The amplitude was so adjusted that the ef-.
GVo 1031 .
~ .
J
_ 21 -factive noise level was slightly less -than 1 volt. r~his noise signal is the sig~al 85 in fig. 12 which figure is a reproduction from a screen posi-tive from an i~stan-t pic-ture o~ -the screen of an oscilloscope on which -the dif~
feren-t signals were made visib:Le. r~he grid of the oscil~
loscope, which has been overdrawn f'or the sake of visibi-li-ty, is composed of squares measuring 1x1 cm. ~he ver-tical sensitivity was adjusted so that 1 cm corresponds with 1 vol-t. ~he time basis was adjus-ted so that 1 cm of deflection in the horizontal direction corresponds with a time interval of 10 ms. ~he signal 86 represents an arbitrarily produced defect pulse 90 with a level slightly greater than 1 volt. ~he signal 87 is the sum of signals 85 and 86. It should be understood tha-t the three sig~als were not simultaneously displayed on the oscilloscope, but that three successive pictures were taken on one film with the vertical positioning of the cathode ray spot so adjusted tha-t for each signal a different vertical position on the screen was taken.
20~ ~he result of a S/N ratio improvement of the signal 87 by 12.79 db as a consequence of additive correlation is illustrated by the curve 88 of fig. 13 whereir~ the verti-cal sensitivity was 1 V/cm, whereas the quite spectacular improvement of 42.14 db by multiplicative correlation is illustrated by the curve 89 wherein the ver-tical sensi-tivi-ty was 10 V/cm.
It will be understood tha-t the invention is not limited to the described embodiment.
It is possible to use the known additive correlation in cascade with the inventive multiplicative correlation to further improve the S/N ratio, and thereby the sensitivity of the system for streaklike defects~
It is further possible to produce the measurement sig-nal Vm in another way~ namely by scanning each section of - GVo1031 5~
the width of the web by only one light source, and by delay;ny the signal produced by said light source over a certain time thereby to compare ;t with the instant signal after such time delay in order to establish an occasional deviation. This technique of simulating the second light source of each pair of sources is disclosed in cletail in our application no. 330,537 referred to hereinbefore~
It is also possible to correlate the output voltages Yp of the photocells after suitable amplification, and then to compare them with a suitable reference signal, "normalized~ as the case may be as lo mentioned in the introduction of the specificat;on.
Other types of delay lines may be used in the correlator, for instance magnetic delay lines, acoustic delay l;nes, charge coupled devices (CCD's) etc.
6V.1031 . "
and that the improvement of the S~N ratio is :
~/N - ~ or, expressed in db : (2) ~/N = 10 log 1 A ~3) ~ he s~stem reaches 99 ~ of its maximum ~/N improve~
men-t after a -time t given by :
t = ~ (4~
As mentioned already in -the in-troduc-tio~ of the speci-fication, a significant improvement of -the S/N ra-tio can only be ob-tained for a gain close to 1, e~g. 0.99, what requires in practice a complete digital lay-out of the system in order to obtain a-stable operation. The time re~
quired for reaching such improvement is not neglectable.
~or instance, for the device illustrated in figs~ 1 to 4 wherein ~ is equal to 0.2 s,t = 100 (from t = ~ 9-).
~ ig. 7 is a diagram illustrating the S/~ improvement for A = O.99,~ = 0.2 s, and t = 100 s. ~he S~N improve-ment amounts to a ra-tio of 14.106. ~he abscissa is a time axis in seconds whereas the ordinate represents V0. The maximum level of the noise component Vn of -the output signal V after 100 s has been arbitrarily illustra-ted at a o ~
level of 1 volt. In fact Vn = Vi~ ~ whereas d l ~ . It may be seen that both the noise signal Vn and the defect signal Vd exponentially increase in the same direction towards a determined limit.
In accordance with the present invention, -the ~/N
ratio of periodically recurring signals is improved by multiplicatively correlating said signals instead of additively correlating them~ ~he basic opera-tion of multi-plic~tive correlation is illustrated in figo 8 wherein the block circuit comprises a multiplicator 70, a delay line 71 with a delay time ~ , and a feedback element 72 with a gain factor A~ It can be shown that the relation between the output voltage ~0 and -the input voltage Vi is expressed GV.1031 :,.
.
~z~
by: [VO = i_O ~ Vi (t - i.t) ~ (5) In consequence, -the sys-tem produces a S~N improve~ent that is infinitely grea-t in -theory. In practice~, said improvemen-t is limited by the maximal level -the signal component Vd can reach in the circuit. ~ommon values for said signal Vd are within 12 and 15 volts in modern inte-grated circuit lay~outs.
A basic condition for the operation of -the described circuit is -that the product of -the input signal Vi with its delayed equivalent is always greater than oneO In case said product is smaller than one, then repea-ted multipli-cation will only reduce the amplitude of the signal instead of increasing it. ~he mentioned condition may be met in one of two ways : either the signal Vi is greater than 1 volt and in that case the gain A may be equal to one, or -the signal Vi is smaller than one volt and in that case the gain A is adjusted so that the produc~t Vi.A is greater than 1.
Another basic condition for the operation of the cir-cuit is that the ratio of the useful signal level Vd to,-the effective noise level Vn of the input voltage Vi be greater than one. Otherwise the system is incapable of'making a dis-tinction between signal and noise. ~he operation of the system is illustrated in the diagram of fig. 9 wherein the abscissa is a time axis in seconds, and the ordinate ,represents the output voltage VO. In case Vi = 1 volt and for A = 1, there is no S/N improvement whatsoever.
In case Vi~ 1, -than it may be seen that as a conse-quence of repeated multiplication the outpu-t voltage VO
increases towards infinity. In practice said vol-tage VO
will reach a limit Vsat which is the saturation vol-tage or the full scale output voltage of the electronic circuitO
In case Vic 1, then the output voltage decreases as a con-sequence of repeated multiplication and tends towards zero.
GV.1031 ', ?,7Z~
Thus it is clear that the signal component Vd of Vi should be greater than 1 whereas the effective noise componen-t Vn f Vi should be smaller -than 1.
~ig. 10 is a block circuit of a mul-tiplicative corre-lator wherein common integrated circuits are u~ed~ ~ mul~i-plier 75 (type 801~ manufactured by Intersil Corp.) is followed by an analogue to digi-tal convertor 76 (type AD 7570 manufactured by ~nalog Devices), a digital 8-bits shift register 77 (-type SY 2533 manufactured by Synertek) ,k,:,",, 10 and a digital to analogue convertor ~ (type AD 559 manufac-tured by Analog Devices). ~he circuit comprises further an analogue adder 79 (type ~A 741 manufactured by ~airchild) for adding to the ou-tput voltage V0 a small voltage Va that may be set by means of a potentiometer 80.
~he purpose of said voltage Va is as follows. At the starting of a correlation cycle, it may occur that, depend-ing on the offset characteristics of the convertor 78, that the output voltage V0 is zero. ~he repeated multipli-cation of V0 by Vi will involve no result since multiplica-20~ tion by a factor zero gives a product that remains zeroO
~ herefore, in order to take account of said possibility that the outpu-t voltage V0 might be zero, a given voltage Va is added to the voltage V0 before the multiplication with Vi occursO Said voltaga Va may be small since its only purpose is -to make sure that Vi will not become multiplied by zero. I~ practice, Va will be chosen between 100 and 200 mV for a full scale output voltage of the device of 10 V. It is clear that the addition of said voltage Va "falsifies" so to say the correlated output signal V0.
It should be understood, however, that this "falsification"
is neglectable as compared by the dis-tortion of the signal Vi as a consequence of the multiplicative correlation. As a matter of fact, any signal Vi that comprises a signal component greater than 1 volt and a noise component smaller GV.1031 :
;' `,j :
ZJ~ ~ 5 than 1 volt, inevitably is amplified after a number of correla-tions to the saturation voltage Vsat of -the device.
Thus, whether the signal componen-t of Vi is, for instance, 1.1 or 1.5 V, the output vol-tage V0 will become within -the shortest time equal to Vsat. r~he described dis~or-~ion of the input signal Vi i5 comple-tely imma-terial in this type of apparatus wherein the o~ly aim is -to increase -the S/~ ratio of' a signal Vi that fulfils the basic conaitions set forth hereinbefore, as rapidly and as much as possible at all.
In case the signal component Vd of a signal Vi is smaller than 1 V, it is yet possible to multiplicatively correlate said signal in amplifyi~g it by a factor A so ~-that the product of Vi.A is greater than 1. ~n amplifier for carryi~g out the mentioned s-tep is illustrated in bro-ken lines by the block 81 in fig~ 10. ~aid block might as well figure between the input terminal 82 of the circuit and the multiplier 75, between the multiplier 75 and the convertor 76, etc., thereby to provide at the lower input 20~ terminal of the multiplier 75 an input signal greater than 1 V.
~ ig. 11 illustrates the device described hereinbefore with reference -to figs. 1 to 4, modified`however~ for mul-tiplicative correlation of the measureme~t signal Vm. ~he part of the drawing situated at the lef-t hand side of the dash and dot line 49 is identical with that of fig. 4 and has therefore been given the same re~erence numerals, except for the amplifiers 44 the gain o~ which has been made adjustable, as indicated by the conventional sign, so that the output measurement signals Vm at terminals 50 may be so adjusted that the noise of the signal at each terminal 50 is just smaller than 1 volt.
The signals Vm form in fact the input signals Vi o~
the multiplicative correlator, the shift register 77 of GVo 1031 .
~ . .
. ~ .
31 ~27 19 ~-which is controlled synchronously with the scanning of the different sections of the web by means of the ou-tput sig~al from the pulse shaper 57. ~he correlator is followed by two comparators 58 and 61 -that have the same function as the comparators disclosed in fig. 4. ~le source 80 has been adjusted for adding a signal Va = 0~2 V
to the output sig~al VO. ~he following da-ta illustrate the operation of the described device that was used for the inspection of a wet coating of silver halide that was applied by means of a slide hopper to a polyethylene terephthalate film having a width of 1770 m.
Scanning amplitude d of the frame 10 : 35 cm ~ffective scanning width c : 21.5 cm -Oscillation period P : 2.4 s 15 ~ffective scanning time ~ : 0.6 s ~umber of optical units : 8 Width of a line of radiant energy on the web : 1 mm ~ength of a line of radiant energ~ : 10 cm 20 Distance between the two lines of a pair of lines : 1 mm Frequency (f) of -the oscillator 43 (that is also the inspection frequency) : 10 kHz Number of signal comparisons during one scan (~xf) : 6000 2 ~umber of signal observations during 5 one scan : 1024 ~umber of signal circulations through a correlator : 5 Corresponding observation time : 12 s S/W improvement obtained by additive correlation according to fig. 4 : 12.79 db (for attenutation A = 0.9) S/N improvement obtained by multiplica-tive correlation according to fig~ 11 0 42.14 db (for 8 bit resolution of the AD
convertor) GV.1031 . .
: ~ ' ,'7 Variation in the output signals for a film fluttering of 1 mm at a frequency above 20 Hz : 0.5 %
Components used :
~amps 35 and 36 : ~D type G~ manufactured by ~'IRE DIVIS:CON
Inter~ational Audio Visual Inc.
Photocells 18 : PIN-LL~ manufactured by Unl-ted De-tector rL~ech-nology Spectral band of measurement : 880 nm wavelength.
It should be noted that the ~ul-tiplici-ty of signal comparisons, namely 6000, -that is carried out by the measuring device with the alternately operated pairs of lamps 95 and 96, is much greater than the number, namel~
1024, of discrete signals into which the output voltage of the measuring device is divided for carrying out the correlation. It can thus be said that finally the re-solving power of the device is 102~ points for each websection being examined. Whereas this number is largely sufficient for many applications, it should be understood it can easily be increased by using instead of pulse strip ~6 another pulse strip capable of producing a larger num-ber of pulses. It is evident that the number of the posi-tions of the shift register 77 should correspondingly be increased. Another point that has to do with the sensiti-vit~ of the device is that -the ~DC convertors 75 and the DAC convertors 78 are 8 bit devices whereby 256 (that is 28) distinct levels of the signals ~i can be transmi-tted~
It is clear that by using convertors with a number of bits 2n, wherein n is greater than 8, the sensitivity for small signal variations can further be increased.
In order to visualize the ~/N ratio improvement that can be ob-tained in accordance with the present invention, the following test was made. A generator for generating a white noise containing frequencies within the range of O to 20 kffz was used to similate the input voltage ~i f a correlator. The amplitude was so adjusted that the ef-.
GVo 1031 .
~ .
J
_ 21 -factive noise level was slightly less -than 1 volt. r~his noise signal is the sig~al 85 in fig. 12 which figure is a reproduction from a screen posi-tive from an i~stan-t pic-ture o~ -the screen of an oscilloscope on which -the dif~
feren-t signals were made visib:Le. r~he grid of the oscil~
loscope, which has been overdrawn f'or the sake of visibi-li-ty, is composed of squares measuring 1x1 cm. ~he ver-tical sensitivity was adjusted so that 1 cm corresponds with 1 vol-t. ~he time basis was adjus-ted so that 1 cm of deflection in the horizontal direction corresponds with a time interval of 10 ms. ~he signal 86 represents an arbitrarily produced defect pulse 90 with a level slightly greater than 1 volt. ~he signal 87 is the sum of signals 85 and 86. It should be understood tha-t the three sig~als were not simultaneously displayed on the oscilloscope, but that three successive pictures were taken on one film with the vertical positioning of the cathode ray spot so adjusted tha-t for each signal a different vertical position on the screen was taken.
20~ ~he result of a S/N ratio improvement of the signal 87 by 12.79 db as a consequence of additive correlation is illustrated by the curve 88 of fig. 13 whereir~ the verti-cal sensitivity was 1 V/cm, whereas the quite spectacular improvement of 42.14 db by multiplicative correlation is illustrated by the curve 89 wherein the ver-tical sensi-tivi-ty was 10 V/cm.
It will be understood tha-t the invention is not limited to the described embodiment.
It is possible to use the known additive correlation in cascade with the inventive multiplicative correlation to further improve the S/N ratio, and thereby the sensitivity of the system for streaklike defects~
It is further possible to produce the measurement sig-nal Vm in another way~ namely by scanning each section of - GVo1031 5~
the width of the web by only one light source, and by delay;ny the signal produced by said light source over a certain time thereby to compare ;t with the instant signal after such time delay in order to establish an occasional deviation. This technique of simulating the second light source of each pair of sources is disclosed in cletail in our application no. 330,537 referred to hereinbefore~
It is also possible to correlate the output voltages Yp of the photocells after suitable amplification, and then to compare them with a suitable reference signal, "normalized~ as the case may be as lo mentioned in the introduction of the specificat;on.
Other types of delay lines may be used in the correlator, for instance magnetic delay lines, acoustic delay l;nes, charge coupled devices (CCD's) etc.
6V.1031 . "
Claims (10)
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:
1. In a method for inspecting at least one longitudinal section of a moving sheet material for streaklike defects, by repeatedly transversely scanning during a period P each such section of the material by means of radiant energy capable of being modulated by said sheet material, and by receiving said modulated energy on at least one photocell thereby to produce at said at least one photocell at least one train of measurement signals V periodically recurring in response with the scanning of the material and which vary in intensity as an indication of the presence of said defects, wherein the S/N ratio of those having a S/N ratio of greater than one is improved by electronic correlation, the improvement where-in said correlation comprises the following steps:
using delay line means for delaying the measurement signals Vm in the train during one scan over a time delay t corresponding with the scanning period P;
using multiplier means for multiplying the measurement signals during a following scan with a factor that is proportional to the corresponding delayed signals, and is greater than 1;
using delay circuit means for delaying the thus multiplied signals over said time delay t and multiplying the measurement signals produced during a subsequent scan with a factor that is proportional to said corresponding previous multiplied signals, and so on for a number of scanning periods.
using delay line means for delaying the measurement signals Vm in the train during one scan over a time delay t corresponding with the scanning period P;
using multiplier means for multiplying the measurement signals during a following scan with a factor that is proportional to the corresponding delayed signals, and is greater than 1;
using delay circuit means for delaying the thus multiplied signals over said time delay t and multiplying the measurement signals produced during a subsequent scan with a factor that is proportional to said corresponding previous multiplied signals, and so on for a number of scanning periods.
2. The method as defined in claim 1, wherein said proportional factor is the product of the previous signals and a constant A which is at least equal to one.
3. The method as defined in claim 1, comprising the step of adding a constant signal Va to said delayed signals, said signal V being comprised between zero and Vm.
4. The method as defined in claim 1, comprising generating a multiplicity of discrete measurement signals during each scanning, comparing each two successive discrete measurement signals with each other and using the difference between each such two successive signals as the signal Vm the S/N
ratio of which must be improved.
ratio of which must be improved.
5. The method as defined in claim 1, wherein the delay line is a digital delay line which is controlled in response to the scanning of the sheet material.
6. The method as defined in claim 1, comprising continuing the repetition of the multiplication of each delayed signal Vm until a predetermined output voltage Vo has been obtained.
7. The method as defined in claim 6, wherein said predetermined value is the saturation voltage of the electronic circuitry for carrying out said correlation.
8. In an apparatus for inspecting a moving sheet material for streak-like defects, comprising cooperating radiant energy means and photocell means arranged for periodically scanning said moving sheet material, means for amplifying the measurement signals Vm that are produced during each scan of the sheet material, and electronic correlation means for improving the S/N ratio of the measurement signals, the improvement wherein said electronic correlation means comprises:
delay lines controlled in response to the scanning of the material for delaying said measurement signals Vm over a delay period equal to the scanning period P;
multipliers for multiplying the delayed signals Vm with corresponding subsequent measurement signals Vm that occur during a following scanning period, and for repeating such multiplication for a number of periods thereby to increase the S/N ratio of those measurement signals having a signal component Vd greater than the effective noise component Vn, and means for evaluating said signals Vd with improved S/N ratio thereby to identify streaklike defects in the sheet material.
delay lines controlled in response to the scanning of the material for delaying said measurement signals Vm over a delay period equal to the scanning period P;
multipliers for multiplying the delayed signals Vm with corresponding subsequent measurement signals Vm that occur during a following scanning period, and for repeating such multiplication for a number of periods thereby to increase the S/N ratio of those measurement signals having a signal component Vd greater than the effective noise component Vn, and means for evaluating said signals Vd with improved S/N ratio thereby to identify streaklike defects in the sheet material.
9. The combination defined in claim 8, wherein said correlation means includes a feedback line between the output of said delay line and the input of said multiplier, and further comprises an adder in said feedback line for adding to the fed back signal an adjustable signal greater than zero.
10. The combination defined in claim 8 having a plurality of separate measurement channels and wherein an amplifier with adjustable gain is provided for each measurement channel, thereby to adjust each measurement signal Vm in such a way that the noise component Vn thereof is smaller than 1 volt but that the signal component Vd is greater than 1 volt.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| GB41.163/78 | 1978-10-19 | ||
| GB7841163 | 1978-10-19 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CA1127259A true CA1127259A (en) | 1982-07-06 |
Family
ID=10500438
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA336,470A Expired CA1127259A (en) | 1978-10-19 | 1979-09-27 | Method and device for inspecting a moving sheet material for streaklike defects |
Country Status (5)
| Country | Link |
|---|---|
| US (1) | US4274748A (en) |
| EP (1) | EP0010791B1 (en) |
| JP (1) | JPS5557137A (en) |
| CA (1) | CA1127259A (en) |
| DE (1) | DE2965385D1 (en) |
Families Citing this family (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4348114A (en) * | 1979-12-07 | 1982-09-07 | Ciba-Geigy Ag | Method of inspecting coated web material to detect the presence of downlines thereon |
| JPS5862252U (en) * | 1981-10-22 | 1983-04-26 | コニカ株式会社 | surface inspection equipment |
| EP0079851A1 (en) * | 1981-11-04 | 1983-05-25 | Ciba-Geigy Ag | Method and device for inspecting sheet material |
| JPS58106409A (en) * | 1981-12-18 | 1983-06-24 | Seiko Instr & Electronics Ltd | Discriminating method for attitude of parts |
| CH663474A5 (en) * | 1984-04-24 | 1987-12-15 | Zellweger Uster Ag | METHOD AND DEVICE FOR THE AUTOMATIC MONITORING OF TISSUE LINES. |
| US4722297A (en) * | 1985-03-25 | 1988-02-02 | The First National Bank | Film coater |
| US4612875A (en) * | 1985-03-25 | 1986-09-23 | Qmi Corporation | Film coater |
| US5118195A (en) * | 1990-09-10 | 1992-06-02 | Rkb Opto-Electrics, Inc. | Area scan camera system for detecting streaks and scratches |
| US5184190A (en) * | 1991-05-28 | 1993-02-02 | Winzen International, Inc. | Method and apparatus for detecting flaws and defects in heat seals |
| US5668887A (en) * | 1992-05-29 | 1997-09-16 | Eastman Kodak Company | Coating density analyzer and method using non-synchronous TDI camera |
| JPH0743313A (en) * | 1993-07-29 | 1995-02-14 | Canon Inc | Foreign matter inspection system and production of semiconductor device using it |
| DE19528519A1 (en) * | 1995-08-03 | 1997-02-06 | Tzn Forschung & Entwicklung | Device for the detection of strip-like surface defects |
| BRPI0720784A2 (en) * | 2007-01-11 | 2014-01-28 | 3M Innovative Properties Co | LONGITUDINAL BLANK SENSOR |
| EP2165162A2 (en) * | 2007-06-19 | 2010-03-24 | 3M Innovative Properties Company | Total internal reflection displacement scale |
| KR101493115B1 (en) * | 2007-06-19 | 2015-02-12 | 쓰리엠 이노베이티브 프로퍼티즈 컴파니 | Systems and methods for indicating the position of a web |
| CN101688794B (en) * | 2007-06-19 | 2012-12-12 | 3M创新有限公司 | Systems and methods for fabricating displacement scales |
| WO2010077592A2 (en) | 2008-12-29 | 2010-07-08 | 3M Innovative Properties Company | Phase-locked web position signal using web fiducials |
| SG172781A1 (en) | 2008-12-30 | 2011-08-29 | 3M Innovative Properties Co | Apparatus and method for making fiducials on a substrate |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US3700876A (en) * | 1970-12-09 | 1972-10-24 | Us Navy | Reduced time delay auto-correlation signal processor |
| US3835332A (en) * | 1973-06-04 | 1974-09-10 | Eastman Kodak Co | Inspection apparatus for detecting defects in a web |
| US3972624A (en) * | 1973-12-20 | 1976-08-03 | Agfa-Gevaert, A.G. | Process and apparatus for detecting longitudinal faults on moving webs of material |
| US4005281A (en) * | 1975-05-14 | 1977-01-25 | E. I. Du Pont De Nemours And Company | Defect identification with normalizing of gain function in optical-electrical inspection |
| GB1526375A (en) * | 1977-03-23 | 1978-09-27 | Ciba Geigy Ag | Method of identifying streaks on a web |
-
1979
- 1979-09-27 CA CA336,470A patent/CA1127259A/en not_active Expired
- 1979-10-04 EP EP79200564A patent/EP0010791B1/en not_active Expired
- 1979-10-04 DE DE7979200564T patent/DE2965385D1/en not_active Expired
- 1979-10-17 US US06/085,711 patent/US4274748A/en not_active Expired - Lifetime
- 1979-10-17 JP JP13470979A patent/JPS5557137A/en active Pending
Also Published As
| Publication number | Publication date |
|---|---|
| DE2965385D1 (en) | 1983-06-16 |
| US4274748A (en) | 1981-06-23 |
| JPS5557137A (en) | 1980-04-26 |
| EP0010791B1 (en) | 1983-05-11 |
| EP0010791A1 (en) | 1980-05-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA1127259A (en) | Method and device for inspecting a moving sheet material for streaklike defects | |
| US3922093A (en) | Device for measuring the roughness of a surface | |
| US3588513A (en) | Method and apparatus for photoelectric inspection of sheet materials | |
| US3043956A (en) | Measuring systems using infra-red radiation | |
| US3958127A (en) | Optical-electrical web inspection system | |
| US3331963A (en) | Apparatus for optically detecting flaws in sheet material | |
| US3026415A (en) | Flaw detector for continuous web | |
| US3228282A (en) | Apparatus for measuring and recording basis weight of paper and the like | |
| US3325649A (en) | Radiation responsive apparatus for monitoring a web | |
| US3612702A (en) | Web defect determination by laser beam irradiation and reflected light examination | |
| US3572099A (en) | Ultrasonic non-destructive testing apparatus and method | |
| US3510664A (en) | Automatic laser beam scanning film flaw detector | |
| US4274747A (en) | Method and device for inspecting a moving sheet material for streaklike defects | |
| US4388832A (en) | Method and apparatus for receiving ultrasonic waves by optical means | |
| US4054377A (en) | Method and apparatus for examining sheet material | |
| US3779649A (en) | Method of and an electro-optical system for inspecting material | |
| EP0211654B1 (en) | Sheet coating thickness measuring apparatus | |
| US4155012A (en) | Discrimination circuit arrangements | |
| JPS6125042A (en) | Surface-defect examining device | |
| JPS6051643B2 (en) | Method and device for detecting work piece surface deformation | |
| US3958128A (en) | System for determining a transversal position of any defect in a traveling sheet material | |
| SE8403402D0 (en) | PROCEDURE AND DEVICE FOR MONITORING OF STRIMOR IN THE MACHINE DIRECTION IN PAPER OR ITS COATING | |
| EP0093890B1 (en) | Apparatus for detecting the irregularities on the surface of a linear material | |
| JPS61254809A (en) | Inferior shape detector | |
| US3800157A (en) | Web inspection apparatus |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| MKEX | Expiry |