CA2022269A1

CA2022269A1 - Method of decoding bar code symbols from partial scans

Info

Publication number: CA2022269A1
Application number: CA002022269A
Authority: CA
Inventors: Theodosios Pavlidis; Joseph Cai; Frederick Schuessler; Jiali Diane Chen
Original assignee: Symbol Technologies LLC
Current assignee: Symbol Technologies LLC
Priority date: 1990-01-05
Filing date: 1990-07-30
Publication date: 1991-07-06
Also published as: EP0436072A3; JPH03204790A; DE69027757T2; AU631232B2; EP0436072A2; AU6005290A; EP0436072B1; US5278398A; DE69027757D1

Abstract

Abstract of the Disclosure A method for decoding bar code symbols in which a bar code reader and associated components scan a bar code symbol produce a representation of the symbol, and attempt to decode that representation in order to produce a character message. The method consists of the steps of storing a first representation of the bar code symbol corresponding to a first scan path through a first portion of said symbol smaller than the entire symbol, and storing a second representation of the bar code symbol corresponding to a second scan path through the symbol. The method then processes said representations in the event a successful decode has not taken place by utilizing the signal from a first scan with the signal from a subsequent scan to determine whether the information from successive scans can be concatenated to form a combined signal.
The process is repeated with different subsequent scans until successful decoding of the combined signal.

Description

Docket No. 067 MET~OD 0~ D2CODING ~R CODg BY~Bo~8 ~RO~ i?AR~IAh ~C~B

R$FERBNCB ~o R~ D CA8B~

This applic~tion is related to U.S. Patent Application Serial Nos. and , filed simultaneously herewith, said applications being assigned to Sy~bol Technologies, Inc.

~AC~GRo~ND OF ~ IN~N~Io~
'.

1. Field of the In~n~en ' This in~ention generally relates to the design of laser scanning system~ for reading bar code sy~bols or similar indicia and, more particularly, to method of decoding a bar code which is ~possibly scanned at~an angle or is poorly printed so that each scan may pro~ide on:ly a partial scan of the entire symbol.

: 1 .

, ~ .
. ' : ....
.. . .

~9,2~
Docket No. 067 2. Description of the Related Ar~
Various optical readers and optical scanning systems have been developed heretofore ~or reading bar code symbol5 appearing on a label or on the surface of an articla. The bar code symbol itself is a coded pattern o~ indicia comprised of a series of bars of various widths spaced apart ~rom one another to bound spaces of various widths, the bars and spaces having di~ferent light-reflectinq characteristics. The readers and scanning systems electro-optically decode the symbol to multiple alphanumerical characters that are intended to be descriptive o~ the article or some characteristic thereo~. Such characters are typically represented in digital form a9 an input to a data processing system for applications in point-of-sale processing, inventory control, and the liXe. Scanning systems of this ge~eral type have ~een disclosed, for example, in U.S. Patent Nos. 4,251,798; 4,360,79~;
4,369,361: 4,387,297; 4,409,470 and 4,460,120, all o~ which have bsen assigned to the sa~e assignee as the instant application.

As disclosed in some o~ the above patents, one embodiment o~
such a scanning system resides, inter alia, in an emitting a light beam frora a hand-held, portable scanning head supportl3d ~y a user, and aimin~ the head, and more particularly, the light beam, at a symbol to be read. The scanner functions by repetitively scanning tha light beam in a line across the symbol.

Docket No. 067 A portion of the reflected light which is reflected off the symbol is detected, and electronic circuitry or software decodes th~ electrical signal into a digital representation of the data represented by the symbol scanned.

A laser scanner includes a light source such as a gas laser or semiconductor laser that generates the light beam. The use of a semiconductor devices as the light source in scanner systems is especially desirable because of their small size, low cost and low power require~ents. The laser beam is optically modified, typically by a lens, to for~ a beam spot of a certain size. It is preferred that the beam spot size be approximately the s~me as the minimum width between regions of dif~erent light re~lectivity, i.e., the bars and spaces of the symbol.

The bar code 5ymbols are formed from bars or elements typically rectangular in shape with a variety of po~sibla widths.
The speci~ic arrangement of elements defines the character r pros2n~ed accordin~ to a set of ~ll~s and de1nitions specified by the code or "s~mbolo~yl' used. The relative size of the bars and spacas is deter~ined by the type o~ coding used, as is the actual size o~ the bars and space~. The nu~ber of characters per inch represented by the bar code symbol is referred to as the density o~: the sy~bol. To encode a desir d sequence o~ character~, a 2 ~13 2 2 ~ ~ ~

Docket No~ 067 collection of element arrangements are concatenated together to form the complete bar code symbol, with each character of the message being repr~-sented by its own corresponding group of elements. In some symbologies a unique "start" and "stop"
character is used to indicate where the bar code begins and ends.
A number of di~ferent bar code symbologilas exist. These s~mbologies include UPC/EAN, Code 39, Code 128, Codabar, and Interleaved 2 of S.

For the purpose of our discussion, characters recognized and defired by a symbology shall be referred to a~ legitimate characters, whilo characters not recognized and defined by that symbology are re~erred to as ille$itimate characters. Thus, an arrangement of elements not decodable by a given symbology corresponds to an illegitimate charactar(s) for that symbology.

In order to increase the amount of data that can be repre~ented or stored on a given amount o~ surface area, sevexal new har code symbologies hava recently been developed. One of these new code standards, Code 49, introduces a "two-dimensional"
concept by stacXing row~ o~ character~ vertically instead o~
extending th~ bars hori~ontally. That is, there are several rows of bar and space patterns, instead of only one row. The structure of Code 49 is described in U.S. Patent No. 4,794,239, which is ~2~
Docket No. 067 herehy incorporated by reference.

A one-dimensional single-line scan, as ordinarily provided by hand-held readers, has disadvantages in reading these two dimensional bar codes; that i~, the reader mus;t be aimed at Pach row, individually. LiXewise, the multiple~scan-line readers produce a number of ~can lines at an angle to one another so these are not suitable for recognizing Code 49 type o~ two-dim~nsional symbols.

In the scanning systems known in the art, the light beam is directed by a lens or similar optical components along a light path toward a target that includes a bar code symbol on the surface.
A scanning component such as a mirror is also disposed in the light path. The scanning component may either sweep th~ beam spot across the symbol and trace a scan line across and past the symbol, or scan the ~ield of view o~ tha scann~r or do both. A scanner also includes a sensor or photodetector. The photodetector has a field of view which extends acros~ and slightly past the symbol and functions to d~tect light reflected ~rom the sy~bol. The analog electrical signal ~rom the photodetector is first typically converted into a pulsz width modulated digital signal, with the widths correspondin~ to the physical widths of the bars and spaces.

.

5:

.

'`, ' ' ' - : . . :
.

Docket No. 067 Such a signal is then decoded according to the speci~ic 5y~bology into a binary repre~entation of the data encoded in the sy~bol, and to the alphanumeric characters so represented.

The decoding process in known scanning systems usually work in the following way. The deco~er rec~ives the pulse with modulated digital signal from the scanner, algorithm attempts to decode the scan. I~ the start and stop characters and the characters between the~ in the scan were decoded successfully and completely, the decoding process terminate and an indicative o~
a successful read (such as an audi~le beep) is proYided to the user. Otherwise, the decoder receives the next scan, performs another decode attempt on that scan, and so on, until a completely decoded scan is achieved or no more scan~ ar~ available.

In the prior art, th0 attempt to decode a scan will fail i~
one or more characters in the sequence (including the start/stop characters) ~ail to be recognized as legitimate characters. The standard decoding algorit~m does not save any results of partial decoding. A a conse~uence a variety o~ different bar codes including long bar code with short height, poorly printed bar codes, or a bar code seriously marred by extraneous markings are impossible or very hard to be decoded by the standard decoding algorithm and repeated attempts must be made by ~he user ~o :
: 6 2 ~ 2 2 r~
Docket No. 067 reorient the scanning line over the s~mbol.

Upon registering a decodQ failure, bar code scanners and decoders known in the prior art di~card the entire scanning data associated with a scan which does not decode, wh:ich for the purpose of our discussion in generic terms we describe as a distorted raw digital image. The. entire scanning process des~ribed above is repeated until a scan is produced which is su~ficiently fre~ from noise or distortion so that the decoder can co~pare and match a legitimate character representation with the raw digital representation derived from the scan.

Such legitimat~ character reprQsentations are typically stored in or accessible to the decoder. However, such a methocl is time consuming, particularly in circumstances where it is difficult to obtain a scan which is not distorted.

Difficulties associated with the inability to read a bar code sy~bol or tha erroneous reading of a bar code 5ymbol ars especially common in applications involving relatively lony bar code messages, i.e., bar code ~ymbols having nu~erou~ element~. ~s the length of the messag~ increases, the number o~ elements and charactars increases, thereby raising the likelihood that distortion will be introduced at so~e point during th~ scan. Accordingly, th~ bar ' , ~
. . '' ~ ~.

~ ~ 2 2 h' ~ ~

Docket No. 067 code reader may scan the sy~bol numerous times before producing a raw digital image suf~iciently free from disto~:ion that it can be decoded.

Further, bar code ~ymbols may contain printing defects which are arranged such that it is unlikely to obtain a single distortion free scan, i.e., valid characters cannot be produc~d and the bar code canno~ be read. Additionally, the bar codes may be printed in a dot matrix style in which case the abili~y to read th~ bar codes depends on the resolution o~ the print. Using conventional methods o~ decoding, such bar code ~ymbols are often unreadable since reading along any single scan line through the symbol would not result in a sequance of valid characters.

Laser scanners are not the only type of optical instrument capable o~ reading ba~ code symbols. Anoth~r type of optical reader is one which is operative being placed by the ussr in direct contact with the symbol to be read. Such readers typicalIy incorporate a non-la~r light source and detectors based upon charge coupled device (CCD) technology in which the size of the detector i~ larger than or substantial~y the same as the symbol to be read. Such scanners ara lightweight and easy to use, but require sub~tantially direct contact or placement of the reader on the symbol ~o enable the symbol to read. Such contact reading is '. ' ' :

Docket No. 067 a preferred mode of operation for some applications or as a matter of personal pr~ference by the user. Contact or near contact reading may also be implemented in an approprial:ely designed laser scanner.

8~RY 0~ T~ IN~TION

Brie~ly, and in general terms, the present invention provides a method for d~coding a ~ymbol, such s~mbol composed of an ordered sequence of characters, each character including a plurality of individual elements o~ different light reflectivity, with a complete symbol bein~ defined as a sequence of a predletermined number o~ characters. More particular, the steps defining the present invention in a fir~t e~bodiment are: ~a) scanning thQ
symbol along a ~irst scanning path and storing a representation of the characters from the scan to form a first sequence; (b) scanning the symbol along another path different from the preceding scanning path and storing a representation o~ the sequence o~
characters ~rom such sc:an to form a second sequence; (c) processing the first seguence and the second seguence by identifying a ~ubsequence of character~ in the first se~uence with a subsequenca of character~ in ~he second se~uence to ~orm an ext~nded sequence. The process proceeds by repeating steps (b) and (c) until all characters in said extended sequence are valid . .

~22s~

Docket No. OS7 characters and constitute the complete symbol.

The invention further provides a method for decodiny bar code symbols in a se~ond embodiment in which a bar code reader and associated components using timing in~ormation associated with the flying spot beam as it scans a ~ar code symbol having a succsssion of linear ~l~ments of different light reflectivity, produce a representation of the symbol and attempt ts decodQ the representation of the ~y~bol in order to produce a legitimate character message containing legitimate character~ which are decodable, all non-d~coda~le characters being illegitimate.

A first representation of the bar code symbol is stored corresponding to a first scan path through a first portion of said s~mbol smal}er than the entire symbol a}ong with ti~ing in~ormation representing the elapsed time between a re~erence time and the moment at which a predetermined bar or space, such as a s~art or stop character o~ the bar code symbol was crosSed by the spot.

A second representation of the bar code symbol is stored corresponding to a second scan path through said symbol along with timing information representing the elapsed time betwèen a second reference time and the moment at which a predete~mined bar or space ~such as a start or stop character of the bar code symbol was ' .

Docket No. 067 crossed by the spot beam.

The timing information is utilized in conjunction with the stored information to reconstitute the complete sy~bol.

The novel features which are considered a~3 characteristic of the invention are set forth in particular in the appended claims.
The invention itself, however, both as to its construGtion and its method of operation, together with additional objects and advantages thereof, will be best understood from the following description of specific embodiments when read'in eonnection with the accompanying drawings.

~rie~ Descript~on oS th~ Dra~

FIGo 1 depicts an illustrative bar code symbol in the Code 39 format, FIG. ~ i~ a table depicting the representations of all characters a~ d~fined by the Code 39;

FIG. 3 is a block diagram depicting components of a con~entional bar code reading device;

.

. . .

~ocket No. 067 FIG. 4A and 4B is a ~low chart depicting a method ~or decoding a bar ccde symbol from multiple scan s:ignals according to the present invention;

FIG. 5A and 5B is a highly simpli~ied diagrammatic representation of scans across a bar code utilized in the present invention;

FIG. 5C is a representation of partial scans through a bar code, showing the timing information used to predict the lo~.ation of the overlapping raw data on successive scans;

FIG. 6 is a highly 5implified diagram representing searching for ths best match decoding to the present invention; and FIG 7 is a highly simpli~ied pictorial representation of a hand held laser scanner in which the present intention may be implemented.

Detaile~ criDtio~ o~ t~ ~ efe~r*~ ~ho~i~e~

The present invention provides a method for the de~ign of :scanning systems for reading bar code symbols or similar indicia and, mor~ particularly, to method of decoding a bar code which is .

.

-Docket No. 067 possibly scanned at an angle or is poorly printled 90 that each scan may provide only a partial scan or representatio~ of the entire symbol.

The present invenkion -also relates to an error correcting method enabling illegitimate characters read in a bar code sy~bol to be iteratively replaced with legitimate characters so as to provide a decoded bar code symbol having only legitimate characters more fully described in copending U.S. Patent Application Serial No. . Specifically, the method iteratively employs other-wise useles~ bar code readings to efficiently produc~ a corrected bar code by replacing illegitimate characters with legitimate characters even only i.~ a portion o~ the symbol has bee.n scanned.
The method includes the steps of storing first and second raw converted representations of first and se~ond scan signals, respectively. The first raw converted representation or image is then corrected, or improved, using the second raw converted representation or imag~ to ~orm a corrected, or improved, first converted representation or imaga. The term "image" as used herein with resp~ct to the invention includes any repres~ntation, modi~ication or derivation o~ the analog scan signal.

The technique according to the present invention in the preferred embodiment may be used with a wide variety of different , ...

2~ 2~
Docket No. 067 symbologie~. However, for purposes of illustration, the invention will be described using Code 39j in which individual bar elements have two possible widths. Referring to FIG 1, a typical Code 39 ~ymbol consists of a leading quiet zone 10, a s;tart character 20, data characters 30, a stop character ~0, and a trailing quiet zone 50.

Referring now to FIGS. 1 and 2, each character of Code 39 is represented by nine elem2nts, five bar~ 60 and four spaces 70.
Further, each element has two possibl.e states, wide or narrow.

While there are 512 unique combinations of nine such binary elements, Code 39 uses only fortyfour to define its character set which is set forth in Figur~ 2. Accordingly, there exists forty-four legitimate characters recognized by Code 39/ and 468 illegitimate characters.

Re~erring to FIG. 3, a conYentional bar code reader co~prises a l~..ght source or las~r 75, a pho~etector 100, a digitizer 130 and a decoder 180. Digitizer 130 comprises an edge detector 140 and a counter 150.

In tha preferred embodim~nt, the laser 75 of the bar code reading device generates an incident laser beam 80 which i swept .

Docket No. 067 across a bar code symbol represented by the dashed line 90. The photodetector lO0 receives a portion of the reflected light 110 from the bar code symbol 90, and generat~s an analog electrical scan signal which varies in time in accordance with changes in the intensity of reflected beam 110.

The analog electrical ~can signal is provided to a converter such as digitizer 130 by way of conductor 120. Digitizer 130, using an edge detector 140 and counter 150, detects when incident laser beam 80 crosses a boundary 145 (see FIG. 1) between any two elements. Since the rate at which the laser beam is swept over the bar code is generally constant, the travel time between two boundaries of an element i~ proportional to the width of the element defined by such boundaries. Accordingly, ko provide a digital representation o~ the width of each element, edge detector 140 controls counter 150 so that the counter may measure the time interval between elemQnt boundaries. Specifically, at a first boundary, counter 150, after being initialized to æero, begins inc.rem~nting an~ cont1nues inçremen~lng as the laser sweeps across the element being measured. At the next boundary, counter 150 receives a stop sign~l by way of line 160 fro~ edge detector 140.
Count~r 150 then stores its present count which is representative of the width of the previously scannPd element. The above cycle is repeated for all elements of the bar code symbol to form a raw , .~ . . . ' ' ' ' .

, ' ; 2 ~ ~
Docket No. 067 digital image of the bar code. The raw digital image may be distorted depending upon the ~uality of the printed image, or the presence o~ dust, dirt, or other damage.

Turning next to Figure 4, there is shown .a ~low chart depicting a method for decoding a bar code symbol from multiple scans according to the present invention. ~ore particularly, the present invention is an extension of conventional decode algoxithms which are set forth in the laft hand portion of Figure 4A. The dashed line 105 corresponds to processing steps eli~inated from a conventional decoding algorithm. The remainder o~ the figure, and in particular the routine 108 marked "stitching", represents improvements in the decoding process added by way o.f the present invention.

Alternatively, in another embodiment of the present invention, the correction or impro~ement described above ~an al~o be per~ormed by first decoding as much ~s possible the raw representation using the character-leYel recoynition ~unction of the standard decode algorithm, and then "~titching" any newly deco~ed characters to the string of characters decoded from previous scan~ The previous scans are u~ ed to form the improved raw images where stitchi.ng ought to occur. Timing informatian, provided by signals from conventional laser scanners such as those ref~renc~d elsewhere in .' ' . .
.
- '' :"'. ' . ' ' : ': : ' 2~3~

Docket No. 067 this specification, may be used to identi~y those points where stitching should occur. This method can replace the "Best Match Searching" algorithm (procedure 122 in Figure 4~) under conditions where the timing information is a good predictor of correct stitching of~et. This condition is met when the scan rate o~ the laser beam is high when compared to any movement of the barcode relative to the laser scanning device. When this alternate method is used, the algorithm can be set to reject excessive movement of the barcode, so as to ensure that no data charact~r~ are dropped vr duplicated when stitching partial scans.

As in conventional decode algorithms, the method of the present invention according to either embodiment first performs a standard decode attempt on ~very scan sent to the decoder ~100, lO1)~ (The r~ference numerals in parenthes~s represent the corresponding steps or block~ shown in the f low charts of Figures 4A and 4B). I~ the result of this decode attempt ~102) i~ a completely decoded scan, no further processing is require~ (103, 104). If however, the r~sult of this decode is a partially decoded scan, the data from the scan is saved for further processing (106).
The stitching algorithm (108) continues to perform decode attempts on scan combinations until a fully decoded symbol is attained (107), or a determination made that such decoding is not attainable (109). Reference is made to U.S. Patent Application 260,692~ f.iled :
.. . . . .
.
. ' ' ' ' , ' ' ' .

h ~ h $ ~
Docket No. 0~7 October 21, 1988, which is hereby incorporated by reference, to describe techniques used to determine whether a scan or partial scan has detected a decodable symbol.

If the standard decode algorithm results :Ln a complete decode (103,104), no processing is performed. However, if the decode is not complete, and this decode includes enough in~ormation such as at least one start~stop character, the scan and the de~oding result are saved ~or furth~r processing (106) by the ~titching algorithm (108). An additional scan is then performed (109), which we a~sume also does not result in a complete decode (103,106) so we again ent~r the stitching algorith~ (10~). Very bad scans (i.e. those that do not contain more than a few characters of decodable data) are automatically discarded.

The description vf tha stitching a}gorithm (108) according to the present invention can be briefly described with re~erence to Figure 4B.

If thc subsequent sc~n is not too bad (~or example, one can find margin~ of the symbol ~120,121) and the data not too short), a best match searching routine ~122) is performed on the previous scan and the cuxre~t one. The best match routine w111 be described in:detail sub~equently. If the ~st match between th~ two ~cans : 18 , .
:
.

~2~
Dock~t No. 067 i~ good enough (123) and has some distance to the second best match, the decode result o~ current scan is sti.tched to the decode result of the previous scan t124).

In the case of the best match is not goQd enough (123,125) or has a too small distance to the second best match, the current scan is discarded, and if too many current scans are discarded successively, the decode result from previous scans are also di~carded, and an initialization i performed (128)~ The algorithm then restarts in the following scan (109). If too many scans are not discarded (129), additional ~cans are acquired (1093.

If the margins are not ~ound ~120), an initialization (128) is also per~ormed and the algorith~ restarts with a new scan ~109).
Similarly, if marg.in3 are found but no previous stored results per~its a suitable match (130,131), an initiali~ation (128) is again performed and the algorith~ restart~ with a new scan ~109).
If a start and stop character is found (131), no initialization is per~ormQd, and a~itional scans are acquired (109).

This prQcess continues until no more scans are avail~ble or a completely d~coding result to be formed either by the standard d~code algorithm ox ~y the stitching algorithm.

: :
: 19 . , . .
- ' ' , .

~22.;~3~

Docket No. 067 By using this new algorithm, improved pexformance of a bar code scanner can be obtained. Figure 5A shows some examples.
Figure 5A shows six scans (labelled by reference numerals 0 to 5) covered the whole bar code, but no single one of them covers the b~r code completely. This case could not be decoded by standard algorithm, but algorithm according to the present invention can decode it by combining partial decoding results. Figure 5B shows another example in which scans O and 1 are not decodable due to poor printing or noise in area P, and scans 2, 3, and 4 al~so not decodable due to similar distortion.s in area Q. The algorithm according to the present invention can also decode it by stitching.

The core part of the new algorithm is shown in the following chart depicted in Figure 4B. When a scan i5 loaded, if two margins are not found, the algorith~ performs an initializatlon.
Otherwise, if no previous decode result recorded, an initial decoding attempt is perfor~ed.

In th proc~ss of an initial attempt, the algorithm tr:ies to decode th~ two end characters. From these decoding re~ults, the scan direction relat~d to the bar code and the end which should be used to start to decode are found. If the partial dècode result include~ at least one start/stop oharacter and is not too short and do not have too many reiections, the partial result and the binary - ~ , ' ' ' .
.: ' , Docket No. 067 scan are saved as represented by the sequence 121, 130, and 131.
Also a relative coordinate is assigned to this initial scan. The start element of the scan is assigned to an initial value if the scan begins with a start/stop charact~r, or the~ end element of the scan is assigned to a stop value if the scan ends with a start/stop character.

If a previous decoding result already exists, a best ~atch searching process is performed (121). Here a binary ~atch rather than a character match is used. Th2 reason i5 that bin~ry matching is more ac~urate than character matching and it is also more difficult to perform character match searching in the case of a tilted scan not covering start and stop characters.

Since the scans ar~ loaded success~ully, the best match occurs in at least one end o~ the scan. Figure 6 shows how the be~t match searching i~ proce~ds.

Figure 6 depicts a represeQtation of four successive scans, with each element o~ a scan repressnted by an asterisX. The algorithm according to the present invention can be illustrated using the first two succes~ive scans shown at the top o~ the Figure. Th~ previous scan trepresented by the first line in Fi~ure 6) is a relatively short scan. The current scan is a longer scan, ~! 1 .

Docket No. 067 and contains more elements. The best match algorithm is then employed to determine which group of elements in the first scan correspond with a group of elements in the second scan, In short, the algorithm ~ttempts to "align" the scans so that th~ additional information from the scan with more in.formation can be "stitched"
to form a combined scan. The process proceeds as follows.

First, the algorithm tries to ~atch for thle left element (for safety, actually use the third element) o~ the previous scan correspondants to the first left element of the current scan (make left edges o~ two scanC even). That also means that the fourth element of the previou~ scan correspond~nts to the second element o~ the current scan, etc. And the right element of the previous scan correspondents to the Xl element o~ the current scan as shown in Figure 6. Then the algorithm trie~ th~ match for tAe left element of the previou3 scan correspondents to the third element of the current scan (the reason o~ not using the second element is that the mat~h is only to try between bar and bar or space and spece). .An~ s~ on, the algorith~ tries Try-limit times. The try times Try-limit is a predict maxi~u~ try times or the try times which ensuro the overlap part o~ two scans aquals to another predict mini~um length, whichever i~ smaller. At the last try above, the right end of the previous scan corre~pondents to the X2 element of the currant scan as shown in Figure 6.

, ` ~22~
Docket No. 067 Second, the algorithm tries to match~for the right element of the previous scan correspondents to the ~irst right element of the current scan (make right edges of two scans even). The try times in this direction is determined by three cases as follows. If X2 is located left of the end of the current scan-(as the second line in Figure 6~ r the try times equals to the dist~nce between X2 and the right end o~ the current scan. If Xl is located right of the end o~ the current scan (as the third line in Figure 6), the try times equals to Try-limit. If Xl and X2 are located different sides of the end of the current scan (as the fourth line in Figure 6), and i~ the diætance between Xl and th~ right end of the current scan is greater than Try-limit, the try times equals to zero, otherwise it equals to Try-limit ~inus that distance.

The algorithm calculake~ match score in each possible match case. In th~ score calculation, the match scores increases when binary elements ~atched, decreases a bigger number wh~n binary elements unmatched, if the highe t match score among all possible match cas~s i~ high enough and the difference between the highest score and the second highest score is also big enough, the best match is satis~ied. A relative coordinate of the current scan is calculated accordin~ to the best match. Then the decode attempt is applied to the current scan under its relative coordina-te.

' 2 ~ ~
Docket No. 067 The current decoding result then is concatenated or to introduce new descriptive term here, "stitched" to the previous result. The rules used in the stit¢hing are shown in ~he follow-ing tabulation:

Cases 1 2 3 4 5 Previous a a r r a Current a r a r b Stitching a a a r In the tabulation above, the reference characters a and b a~ove represent legal character results from decoding, r represents rejection. If both decode results are legal and equal, then t~e ~titching result is the same characterO If one o~ them i5 legal character and the other is a rejection, the stitching result is the legal character. I~ any contradiction occurs, like in the case 5 above, that is both of decode results are legal but different, then the algorith~ discards the current and previous decode results, performs an initialization (I above stands for Initialization) and restarts in thP following scan, a~ ~hown in 124, 126, 12~ sequence in Figure 4(b).

A completion check is perorm~d afterwards i~ no contradiction occurs in the process of decode result stitching. The completion means that both start/stop characters ar~ found in the combined Docket No. 067 decode result and there are no r~jected characters exist between these start~stop characters. The algorithm terminates in a succeed in the case of completion, goes on next scan otherwise.

If the best match is not satisfied, the curr~nt scan is discarded as shown in 125 of Figur~ 4(b). The next scan will be used to find the best match with the previous scan. If there are too many scans discarded successively (129), the previous scan and its decode result are discaxded and an initialization is performed (128).

The algorithm attempts to match parts o~ the two str:ings first starting ~rom the le~t and then starting from the right. The procedure involves two parameters: "OFFSET," which is an of~set ~rom the start of th~ previous scan and "~ IF~'q which the maximum shift attempted. Although we used Figure 6 above t3 illustrate the proce~, another de~cription is to describe the scan matching process by considering sequence of elements as shown in the text belaw. The upper case letters of the alphabet (A,B,C) represent bar elements, and the lower case elements (a,b,c) space elementsO Consider two scans, labelled "previous scan" and "current scan"~ Assume the data detect~d rom the `scans is as shown below:

~ 25 ~2~2~

ocket No. 067 CbCeA~BbC~C~Ab previous scan /
AeBbB~CbCbAb~b~b current scan To illustrate the algorithm, we start with the element at distanca OF~8~ from the left of the previous scan and we attempt to match it with the leftmost ele~ent of the current scan. If we select OF~T equal to three, then we compare the third element from the left o~ the previous scan ~C in the example) to the leftmost element of the current scan (A in the axample~. While in the case of lett~rs as shown above matching is absolute (two characters either match or do not) in the cas~ of numberR matching is relative: for example, 20 may be matched by either 18 or 21 with 21 offering a hetter match than 18. Thus when compare scans we give a score for the matching (see below). If the initial match does not have a Yery high score, then we try to match to the next element of the current 5can ~it will be B in the example because we can compare bar~ only to bars). This process is repeated ~$-8~ ' times. AftPr ~hs matching o~ the twc elements we attempt to match also subs~quent ones. In the above example the subscans matched are Ao~b~CbCb~b starting from the fifth position in the pr~vious scan and in the first position in the current scan~ ~he rightmost matched element of the current scan is th~ twelfth one:
X1 - 12~ Since it is possible to have more than one match the : 26 --, . .. . ~ :
' 2~22~

Docket No. 067 rightmost position where such a match i5 possible is denoted by x2 and it is always great~r then X1. It may also happ~n that S1 and X2 are past the end of the current scan by ext:rapolation: if the previous scan is longer than the current scan and it matches up to the end of the current scan, then X1 and X2 are defined to correspond to the end of the current scan.

The next step is to match the element of the previouC; scan which is OFF8~T position from the right to the rightmost el ment of the current scan and if that fails to shift to the left in th~
previous scan, and so forth up to a certain number of attempts ~
ATTE~PTS. N-AT~B~P~8 is dete~mined as following: (a) If X2 is less than the length of the current scan N~ MP~8 2quals MAS-8~I~T.
(b) If Xl and X2 are located at dif~erent sides of the end of the current scan then we set W-A~T~MP~8 equal to ~S-~IF~ minus the difference X2-Xl provide that the resulting value is positive.
Otherwise N-ATT~T8 is set to z2ro. Note that no matching from the right is po~sibl~ in the above example.

In the typical environment, the motion of ~he barcode relative to the scanning device is 910w or even negligibl~ compar~d to the rate at which the laser spo~ is scanned repeatedly back and forth acro~s the barcode. Furthermors, the design of laser scanners such as those mention~d elsewhare in this application is such that a .

.

~' , ~2~

Docket No. 067 electrical signal designated as the Start of Scan ~SOS) signal is provided that reverses polarity in a manner that i5 synchronized in time with reversals of the lasex scanning clirection. In particular, if, for illustration, the bar code exhibits no motion relative to the scanning device, and if the SOS signal changes from low to high when a left~to-right scan begins, then it will ~ollow that the elapsed time, measured ~rom the low-to-high change on SOS
until the time the laser spot crosses the first bar o~ the bar code, wîll be the same on every left-to-right scan. Any difference in these elapsed time measurements can there~ore be ascribed to motion of the barcode relative to the scanner, said motion creating a pattern ~uch as that shown in Figure 5C.

A stitching algorithm that relies on the above characteristics will, by using timing predictio~s, limit the search for a best match to a small percentage of the raw data; the algorithm Call also control how far away from the predicted time it will allow a match, and thus prevent mis~ed character~ due to excessive motion. This approach requires that, in order to begin stitching, two successive scans in th~ sa~e direction pass over the same recognizable section o~ the bax code (e.g. the start or stop character). This allows the time difference (due to relative motion o~ the barcode) to be established, so that the stitching region can be quickly and accurately found on subseq~ent scans. This is best explained by , ~ . ~'' . . .

' t~ ~ ~
Docket No. 067 an example, which for brevity will use the simplest (though by no means the only) case, depicted in Figure SC, where, due to motion of the barcode and/or of the scanner, two successive left-to-right scans (labelled SO and S2) both cross the start character ("*"), but no one scan crosses the entir~ bar code. In this simple case, the right-to-left scans (S1, S3, S5, etcO) do not need to be processed, and for clarity are not shown on Figure 5C. The algorithm for this case can be described as follow~O

After initializing the process when the laser is ~irst turned on, acquire scans until two successive scans in the same direction include a start character, as determined by using the standard decode algorithm (scans SO and S2 in Figure 5C). For each scan with a recognized start character, measure the time from the Start of Scan change to when the laser crossed the first bar of the start character (SOS to tO on sc~n SO, and SOS to t2 on scan S2).
Calculate the time difference for the two measurements (d~ltaO =
tO - t2), which reprss~nts the relative motion of ~he barcode and the scanning device. For S2, also calculate the time from SOS to the start of the la~t ~uccessfully-decoded character ("2") of S2 (t20). ~ased on this information, calculate a predicted time (for the next scan S4) between the next SOS change to the start of character "2" on scan S4 (t4predict d - t20 - deltaO). After the raw data ~or scan S4 is loaded, search for the character "2~' (using 2~22~
Docket No. 067 the standard charact~r-decode algorithm) in a small region of raw date c~ntered ~round the tim~ of t4predicted. If the desired data character is found near its predictad time, then continue decoding additional charactars until either the character-decode algorithm ~ails (as it would on scans S4 and S6~ or until end of the barcode is successfully reached (as on scan S8). on incomplete scans, repeat the prediction process to locate the last good character of scan S(n) within the raw da~a of scan 5(n+2~.

To summarize, the second embodil~ent of the present invention provides a method for decoding bar coda s~mbO13 in which a bar code reader and associated components using a flying spot beam scans a bar code symbol having a succes~ion of linear elements o~ different light reflectivity, produce a representation of the symbol and attempt to d~code the representation of the symbol in order to produce a legitimat~ character message containing legitimate characters which are decodable, all non-decodable characters being illegitimate. More particularly, the improve~ent includes the steps of storing a ~irst rapresentation of the bar code symbol corresponding to a first ~can path through a first portion of the 5ymbOl smaller than the entire sy~bol, the representation including information measuring the elapsed time between (i) a reference tima that has a fixed relationship to the time o~ the start of said scan and (ii~ the moment at which a predetermined bar or space of the : 30 . ~ .
':

. . '. ~

Docket No. 067 bar code sy~bol, such as the first ~ar of its start or stop character, was crossed by the laser spot, such elapsed time to be referred to below as t1~ A second representation of the bar code symbol is stored corresponding to a second scan path of said symbol, the representation includin~ information measuring the elapsed time between ti) a re~erence time that has a fixed relationship to the time of the start of said s an and (ii) the moment at which a bar or space, recognizably the same as the bar or space used in the first scan, was crossed by the laser spot, such elapsed time to be referred to below as t2. The di~ference between tl and t2 is cal~ulated and defined as t3, to represent the relative motion between the scanner and the bar code symbol.

The second stored representation is used to measure the elapsed time between the reference time related to the start of the second scan and the time that the laser spot bea~ crosse~ the last decodable character during the second scan so that a string of decodable characters is defined, the length of said string being smallex than the length of the bar code symbol, the elapsed time bei~g defined as t4.

One adds t3 and t4 to compute a time tS~ which, for the next scan to be acquired, represents th elapsed time between the re~erence ti:- ~rom the start ot the thircl scan and the tlre that .

~2~
Docket No. 067 the laser spot will begin to cross the same sequence of ch~racters in the substring used in the s2cond scan.

Ont then uses tS as the reference time to locate and decvde the stored representation of a new character in the next scan, so that the stored representations can be combined to form an improved representation.

Th~ present invention may be implemented in a hand--held, laser-scanning, bar code reader unit such as illustratad in Figure 7. This hand-held device of Figura 7 is generally o~ the style disclosed in U.S. Patent ~,760,248, issued to Swartz et al, assigned to Symbol Technologies, Inc., and also ~imila.r to the configuration of a bar code reader commercially available ~s part ...
number LS 8100II from Symbol Technologies, Inc. Alternatively, or in addition, ~eatures o~ U.S. Patent 4,3a7,zg7 issued to Swartz et al, or U.S. Patent 4,409,470 issued to Shepard et al, both such patents assigned to Symbol Technologies, Inc., may be employed in constructing the bar code reader unit of Figure 7. These patents 4,760,2~8, ~,3~7,297 and 4,409,470 are incorporated herein by re~erence. A outgoing light beam 151 is generated in the reader 100, usually ~y a laser diode or the like, and directed to impinge upon a bar code sym~ol a few inches from the front of the reader unit. I'he outgoing beam 151 is scanned in a fixed linear pattern, : : 32 ,.

2~2~

Docket No. 067 and the user positions the hand-held unit so this scan pattern traverses the symbol to be read. Reflected light 152 from the symbol is detected by a light-responsive device 146 in the reader unit, producing serial electrical signals to be process~d for identiPying the bar code. The reader unit 100 is a gun shaped device, having a pi5tol-grip type of handle 153 and movabl~ trigger 154 is employed to allow the user to activate the light b~am 151 and detector circuitry when pointed at the symbol to b~ read, thereby saving battery life if the unit is self-pow~red. A light-weight plastic housing 155 contains the laser light source, th~
detector 146, the optics and slgnal processing circuitry, and tha CPU 140 as well as a battery 162. A light-transmis~ive window 156 in the ~ront end of the housing 155 allows the ou~going light beam 151 to exit and the incoming reflected light 152 to enter. The reader 100 is designed to be aimed at a bar code symbol by the user from a position where the reader 100 is spaced from the symbol, i.e., not touching the symbol or moving across the symbol.
Typically, thi~ typa o~ hand-held bar code read~r i8 specified to operate in the range of perhaps several inches.

As seen in Figure 7, a su.itable lens 157 (or multiple lens system) is used to colli~ate and focus the scanned beam into the bar code symbol at an a~propriate reference plane, and this same lens 157 may be used to focus the reflected light 152. A light Docket No. 067 source 158 such as a semiconductor laser diode i5 positioned to introduce a light beam into the axis of t.he lens 157 by a partially-silvered mirror and other lense~ or beam-shaping structure as needed, along with an oscillating mirrQr 159 which is attached to a scanning motor 160 activated when the trigger 154 is pulled. If the light produced by the source 15~ is not visible, an aiming light may be included in the optical system, again employing a partially-~ilvered mirror to introduc~ the bea~l into the light path coaxially with the lens 157. The aiming light~ if needed, produces a visible-light spot which is scanned just like the laser beam; the user employs this visible light to aim the reader unit at the symbol before pulling the trigger 154.

In real applications, either the length of bar code should be fixed, or an additional character should be placed in the bar code to indicate its length. If not, mi9decodings may occur.

I~ the length of the bar code is fixed, the performance of decoding using stitc~ing might be. bettPr than that of decoding using a complete scant because the reject rate and the misdecode rate is 12sa in ~o~t goo~ quality bar code (in cases that the ~uality o~ the bar code is very very poor, the misdecode rate may be yreater, but never greater then 2K times, ~ is the number of the scans used ~or stitching).

Docket No. 067 Although the present invention has been described with respect to linear or single line bar codes, it is not limited to such embodiments, but may also be applicable to stacked or two dimensional bar codes such as Code 49 and similar symbologies. It is conceivable that the method of the present invention may also find application for use with various machine vision or optical character recognition applications in which in~ormation is derived from other types of indicia such as characters or ~rom the surface characteristics of the article being scanned.

In all o~ the various e~bodiments, the elements of the scanner may be assembled into a very compact package that allows the scanner to be fabricated as a single printed circuit board or integral module. Such a module can interchangeably be used as the laser scanning element for a variety of different types of data acquisition system~. For example, the module may be alternately used in a hand-held scanner, a table top scanner attached to a flexible arm or mounting extending over the s~rface of the table or attached to the underside of the table top, or mounted as a subcomponent or subassembly of a more sophisticated data acquisition system.

The module would advantageously comprise a laser/optics subassembly =ounted on d support, a scanning ele=ent such as a .

Docket No. 067 rotating or reciprocating ~irror, and a photod~teckor component.
Control or data lines associatsd with such components may b~
connected to an Plectrical connector mounted on the edge or external surface of the module to enable the module to be electrically connected to a matinq connector associated with other elem~nts of data acquisition system.

An individual module ~ay have specific scanning or decoding characteri~tics associated with it, e.g. operability at a certain working distance, or operability with a speci~ic symbology or printing ~ensity~ The characteristic~ may also be de~ined through the manual setting o~ control switches associated with the module.
The user may also adapt the data acquisition system to scan di~ferent types of article~ or the system may be adapted for di~erent applications by interchanging modules on the data acquisition system throu~h the use of the simple electrical connector.

The scanning modul~ described above may also ~e implemented within a self-contained data acquisition system including one or more such components as keyboard, display, data storage, application software, and data ~ases. Such a system ~ay also include a co~munications interface to permit the data acquisition system to ommunicate with other components of a local ~rea network 22~
Docket No. 067 or with the telephone exchange network, either thr3ugh a modem or an ISDN interface, or by low power radio broadcast from the portable terminal to a stationary receiver.

It will be understood that each of the features described above, or two or more together, may find a useful application in other typ~s of scanners and bar code readers di~ering from the types described above.

While the invention has been illustrated and described as embodied in a method of decoding bar code syMbols from partial scans, it is not intended to be limited to the details shown, since various modifications and ~tructural change3 may be made without departing in any way from the spirit o~ the present invention.

For example, although the present invention is described in its pre~erred embodiment a~ relating to the decoding o~ bar code s~mbols in which the detected representation of a character is either decoda~le or not decodable according to ~ome single acceptability criteria, the invention may also be applied to detection of: indicia having a multiple threshold level of acceptability ranging ~rom an unacceptable or "distorted"

reprèsentation of a image to an acceptable or l'clear"
representation~of the image. The iterative process ~ay bQ defined , :

Docket No. 067 as attempting to correct portions o~ the scanned image from relatively distorted to relatively clear in a sequenc~ of steps.

Without further ana].ysis, the foregoing will 50 fully reYeal the gist of th~ present invention that othars can readily adapt it for various applications without omitting features that, from the standpoint of prior art, fairly con~;titute essential characteristics of the generic of specific aspects of this invention and, thersfore, such adaption shol'ld and are int:ended to be comprehended within the meaning and ran~e of eguivalence o~
the ~ollowing claims.

-.
.' ~: . . .

Claims

1. A method for decoding a bar code symbol composed of an ordered sequence of characters, each character including a plurality of individual elements of different light reflectivity.

(a) scanning said symbol along a first scanning path and storing a representation of said characters from the scan to form a first sequence;

(b) scanning said symbol along another path different from the preceding scanning path and storing a representation of the subsequent sequence of characters from such scan to form a second sequence;

(c) processing said first sequence and said second sequence by identifying a sequence of characters in the trailing position of one of said sequences with a sequence of characters in a starting position of the other of said second sequences to form a new first sequence; and (d) repeating steps (b) and (c) until all characters in said symbol are stored.

Docket No. 067

2. A method as defined in claim 1, wherein said elements extend in parallel on a surface so as to form a linear pattern of bars and spaces.

3. A method as defined in claim 1, wherein said sequence of characters include a unique start character as the first character and a unique stop character as the last character, as considered in the direction of scanning.

4. A method as defined in claim 1, wherein said step of scanning comprises directing a laser beam in a scan pattern of mutually parallel scan lines.

5. A method for decoding a symbol composed of an ordered sequence of characters, each character including a plurality of individual elements of different light reflectivity with a set of predetermined patterns of elements representing valid characters, while all other patterns represent invalid characters, comprising the steps of:

(a) scanning said symbol along a first scanning path and storing a representation of said characters from the scan to form a first sequence;

Docket No. 067 (b) scanning said symbol along another scanning path different from the preceding scanning path and storing a representation of the subsequent sequence of characters from such scan to form a subsequent sequence;

(c) processing said first sequence and said subsequent sequences by joining characters in said first sequence with valid characters in said subsequent sequence to form a combined sequence;
and (d) repeating steps (b) through (c) until all characters in said combined sequence are valid characters and represent the entire symbol.

6. A method for reading indicia having portions of different light reflectivity such as bar code symbols by a scanner comprising the steps of:

(a) scanning a target containing a bar code symbol;

(b) detecting at least a portion of light of variable intensity reflected off the target with a sensor and generating an electrical signal corresponding to data represented by the symbol;

Docket No. 067 (c) processing said electrical signal to determine whether the reflected light of variable intensity is indicative of the presence of a predetermined indicia pattern such as a bar code symbol and to terminate said scanning upon the successful decoding thereof;

(d) further processing said electrical signal in the event a successful decode has not taken place by utilizing the signal from a first scan with the signal from a subsequent scan to determine whether the information from successive scans can be concatenated to form a combined signal and indicative of a presence of a predetermined indicia pattern such as a bar code symbol; and (e) repeating step (d) with a different subsequent scan until successful decoding of said combined signal.

7. A method as defined in claim 6, wherein said step of scanning comprises directing a laser beam in a scan pattern of mutually parallel scan lines, and said step of scanning is terminated upon successful decoding.

8. A method as defined in claim 6, further comprising the step of storing said electric signals in a memory array by addressing the rows and columns of said memory array in a sequence Docket No. 067 of addresses that represents a linear traversal of the field of view corresponding to scanning.

9. A method as defined in claim 8 wherein the scanner is contained in a hand-held unit which includes a manually-activated trigger, and the method further includes the step of manually activating said trigger by the user to initiate said step of scanning said target.

10. A method of decoding a bar code symbol that may contain distortion, said symbol including a sequence of characters, each character including a plurality of individual elements of different light reflectivity, with a set of predetermined patterns of elements representing legitimate characters, while all other patterns represent illegitimate characters, comprising the steps of:

(a) storing a first representation of said bar code symbol derived from scanning through said bar code symbol;

(b) storing a second representation of said bar cods symbol derived from scanning through only a portion of said bar code symbol; and Docket No. 067 (c) improving said first representation using said second representation to form an improved representation.

11. The method of claim lo wherein said step of improving said first representation comprises the steps of:

(a) replacing at least one element of a distorted character of said first representation with at least one element of a corresponding character of said second representation to form said improved representation.

12. The method of claim 11 wherein said step of replacing results in said improved representation containing only legitimate characters.

13. The method of claim 11 further comprising, if the entire bar code symbol is not decodable using said improved representation, the step of:

(a) attempting to improve, by element replacement, said improved representation using a third representation of said bar code symbol derived from scanning through at least a portion of said bar code symbol using said improved representation to form a further improved representation.

Docket No. 067

14. The method of claim 13 further comprising, if said step of attempting to improve said second representation was successful, the step of:

(a) attempting to decode the entire bar code symbol using said further improved representation.

15. In a method for decoding bar code symbols in which a bar code reader and associated components can a bar code symbol having a succession of linear elements of different light reflectivity, produce a representation of the symbol and attempt to decode the representation of the symbol in order to produce a legitimate character message containing legitimate characters which are decodable, all non-decodable characters being illegitimate, the improvement comprising the steps of:

(a) storing a first representation of the bar code symbol corresponding to a first scan path through a first portion of said symbol smaller than the entire symbol;

(b) storing a second representation of said bar code symbol corresponding to a second scan path through said symbol; and (c) replacing at least one element of an illegitimate Docket No. 067 character of said first representation with at least one element of a corresponding character of said second representation to form an improved representation.

16. The improved method as defined in claim 15, wherein said second scan path is smaller than the entire symbol.

17. The improved method as defined in claim 15, further comprising the step of utilizing representations from scans through a subsequent scan path through at least a portion of said symbol to form a further improved representation.

18. The improved method 17 further comprising the steps of repeatedly redefining said improved representation and said further improved representation until either the decode is successful or no further scans are available.

19. In a method for decoding bar code symbols in which a bar code reader and associated components using a flying spot beam scans a bar code symbol having a succession of linear elements of different light reflectivity, produce a representation of the symbol and attempt to decode the representation of the symbol in order to produce a legitimate character message containing legitimate characters which are decodable, all non-decodable Docket No. 067 characters being illegitimate, the improvement comprising the steps of:

(a) storing a first representation of the bar code symbol corresponding to a first scan path through a first portion of said symbol smaller than the entire symbol, said representation including information measuring the elapsed time between (i) a reference time that has a fixed relationship to the time of the start of said scan and (ii) the moment at which a predetermined bar or space of the bar code symbol, such as the first bar of its start or stop character, was crossed by the laser spot, such elapsed time to be referred to below as t1;

(b) storing a second representation of the bar code symbol corresponding to a second scan path of said symbol, said representation including information measuring the elapsed time between (i) a reference time that has a fixed relationship to the time of the start of said scan and (ii) the moment at which a bar or space, recognizably the same as the bar or space used in the first scan, was crossed by the laser spot, such elapsed time to be referred to below as t2;

(c) calculating the difference between t1 and t2 defined as t3, to represent the relative motion between the scanner and the Docket No. 067 bar code symbol;

(d) using the second stored representation, measuring the elapsed time between the reference time related to the start of the second scan and the time that the laser spot beam crosses the last decodable character during the second scan so that a string of decodable characters is defined, the length of said string being smaller than the length of the bar code symbol, said elapsed time being defined as t4;

(e) adding t3 and t4, computed a time t5, which, for the next scan to be acquired, represents the elapsed time between the reference time from the start of the third scan and the time that the laser spot will begin to cross the same sequence of characters in the substring used in the second scan;

(f) using t5 as the reference time to locate and decode the stored representation of a new character in the next scan, so that the stored representations can be combined to form an improved representation.

20. The improved method as defined in claim 19, further comprising the step of utilizing representations from scans through a subsequent scan path through at least a portion of said symbol Docket No. 067 to form a further improved representation.

21. The improved method as defined in claim 19 further comprising the steps of repeatedly redefining said improved representation and said further improved representation until either the decode is successful or no further scans are available.