DE202012102113U1 - Optoelectronic code reader - Google Patents

Abstract

Optoelectronic code reader (10) with a light receiving element (20) for converting received light into image data and with an evaluation unit (22) which is designed to find code areas (14, 40) in the image data and code information from the code areas (14, 40) read out, characterized in that the evaluation unit (22) is designed to find a first code area (14, 40) and a second code area (14, 40) in the image data, the image data from the first code area (14, 40) and the second code area (14, 40) on the assumption that codes (40) which are redundant to one another are contained therein to form a common code (46) and then to decode the common code (46) in order to obtain the code information.

Description

The invention relates to an optoelectronic code reader s according to the preamble of claim 1.

The most common code readers are bar code scanners which scan a bar code or bar code with a laser read beam across the code. They are often used in supermarket cash registers, automatic parcel identification, mail sorting or baggage handling in airports and other logistics applications. Other applications include quality or parts controls in manufacturing processes. With the advancement of digital camera technology, barcode scanners are increasingly replaced by camera-based code readers. Instead of scanning code areas, a camera-based code reader uses a CCD chip to take pictures of the objects with the codes on them. Camera-based code readers can easily cope with code types other than one-dimensional barcodes, which, like a matrix code, are also two-dimensional and provide more information.

Conventionally, the regions of potential code candidates segmented by the image processing are transferred sequentially to a decoder stage. The decoder stage converts this region via a sampling grid or a projection into a binarized two-dimensional bit matrix, from which in turn the original message contained in the symbol is extracted or decoded.

The image data of the codes to be read are superimposed during real operation of faults. By clever choice of coding errors can be corrected to a certain extent when reading out. If the code area is severely destroyed, the maximum error correction capability of the symbol can be exceeded. Such codes are therefore completely unreadable. An example of such disturbances are so-called fingers in the code regions, ie printed conductors on wafers.

In many cases, the code exists multiple times in an image captured by the camera or in a sequence of images. For example, the same code is printed multiple times on a printed circuit board or other object and thereby multiple times in the image. In another important application group, the code-bearing objects are in motion relative to the code reader and are recorded multiple times. This happens, for example, in a so-called reading tunnel, in which objects are conveyed past one or more permanently mounted code readers. The code is then contained in several consecutively recorded images with an offset corresponding to the relative movement.

In the prior art, however, the decoding attempts always take place without consideration of the code candidate regions which have already been evaluated or are yet to be segmented. The basically existing code redundancy is thus not recognized and at least not used.

The EP 1 096 416 B1 discloses a code reconstruction in one-dimensional codes. It is conceivable that the scan lines are at an unfavorable angle to the one-dimensional code, so that no scan line traverses the entire code. Then complete detection of the code information is achieved by overlapping stitching of several individual scans. In the overlap areas, where multiple subsets of identical code information are available, an averaging or majority decision can be made to determine which item of information to use for decoding. Since this overlapping evaluation is performed only at interfaces of the individual scans, the correction of errors in the detection of the codes is insufficient. In addition, the problem of partial acquisitions does not arise for camera-based code readers and two-dimensional codes because no angle dependence is given here, or the relative orientation of the codes in the image data can be corrected by image processing.

It is therefore an object of the invention to enable a successful code reading even with heavily destructive codes.

This object is achieved by an optoelectronic code reader according to claim 1. The invention is based on the basic idea that the code content can still be read despite the destruction of the code, which lies above the capacity of the error corrections of the decoder, with the aid of the existing code redundancy. For this purpose, the evaluation unit searches in an image or in an image sequence for decoding the code in a first code area for a further second code area from which it can be assumed on the basis of the position or chronological sequence that it is a redundant code. The image data in the first and second code areas are then combined with each other by utilizing the redundancy to compensate for defects. Only the resulting common code, in which the flaws are reduced or even eliminated, is then decoded.

The invention has the advantage that, by exploiting code redundancy, the read rate of a code reader is increased, especially in the case of difficult image material or when reading in motion can if the same code content exists multiple times in an image or image sequence. The respective correct code contents can be composed by combination, which can be implemented in particular in the form of simplest logical operations. This also makes it possible to detect codes that could not be read in the conventional sequential single-decoding method due to the limitations of the error correction method of a decoder.

The code reader is preferably designed as a camera-based code reader, wherein the light receiving element is a matrix or line-shaped image sensor. In this case, the image sensor has, for example, a multiplicity of light-receiving elements or pixels in CCD or CMOS technology arranged to form a line or a matrix. Although image data can basically also be acquired with the aid of a scanning beam of a code scanner, it is much easier to generate an image that is accessible to digital processing via an image sensor.

The evaluation unit is preferably designed to find code areas with two-dimensional codes and to decode two-dimensional codes. In a two-dimensional code, much more information can be accommodated on the same surface than in a barcode. There are a number of standards known for two-dimensional codes.

The image data preferably has an image with several code areas. This applies to applications in which the objects carry the same code redundantly multiple times. This is not uncommon anyway, to support readings from different perspectives.

The image data comprises a plurality of images with code regions in a preferred alternative. The redundant codes are thus distributed to different images. Corresponding image sequences arise in particular when the same code-carrying object is recorded several times in different stages of motion, such as during code detection on a conveyor belt or in a reading tunnel. The redundant codes have long existed in such applications, but are usually not or at least used only for independent sequential decoding attempts without exploiting the redundancy over a combination of multiple images.

The evaluation unit is preferably designed to binarize the image data and in particular the code regions. This creates a black and white image, which is sufficient for reading most types of code. Disturbances in greyscale images due to different brightnesses are eliminated by the binarization. In addition, the bitmatrix resulting from binarization can be more easily manipulated with simple bit-logic operations. Intelligent binarization algorithms are able to largely avoid misregistration of dark and light pixels.

The evaluation unit is preferably designed to negate the code areas. Thus, after the binarization, the bit values are interchanged so that bright pixels become dark pixels and vice versa. For the combination of code areas, the negated picture may be the better starting point than the original picture.

The evaluation unit is preferably designed to combine code areas or negated code areas by means of logical AND or by means of logical OR to form a common code. The processing is simple and extremely fast, but still capable of reconstructing those due to a destruction of the missing information of one code area from the other code area.

• – durch logische UND-Verknüpfung der binarisierten Bilddaten des ersten Codebereichs und des zweiten Codebereichs
• – durch logische ODER-Verknüpfung der binarisierten Bilddaten des ersten Codebereichs und des zweiten Codebereichs
• – durch logische UND-Verknüpfung der negierten binarisierten Bilddaten des ersten Codebereichs und des zweiten Codebereichs
• – durch logische ODER-Verknüpfung der negierten binarisierten Bilddaten des ersten Codebereichs und des zweiten Codebereichs.

The evaluation unit is preferably designed to combine the first code area and the second code area by means of at least one of four bit-logical operations, namely:

• By logically ANDing the binarized image data of the first code area and the second code area
• By logically ORing the binarized image data of the first code area and the second code area
• By logically ANDing the negated binarized image data of the first code area and the second code area
• By logically ORing the negated binarized image data of the first code area and the second code area.

All these operations are fast and implementable on simple hardware. Depending on the characteristics of the codes and the destructions, a specific selection of the mentioned bit-logical operations is particularly suitable for the reconstruction of the code information from the redundant code areas.

The evaluation unit is preferably designed to execute all four bit-logic operations and to decode the respectively resulting common codes. The bit-logical operations can be carried out parallel to each other or one after the other. By determining the result of all these four bit-logic operations, no particular polarity characteristics of the codes and destruction are assumed. The polarity of the code means that a darker one Code is searched on a light background or vice versa, and according to polarity of destruction, whether it is caused by a dark or a bright area, so for example by gaps in the image or override areas. In principle, it is sufficient if the decoding succeeds after one of the mentioned bit-logical operations. It is then possible to dispense with the further bit-logical operations with subsequent decoding attempt. Alternatively, any one of the successful decodes is used as the result, or a mutual plausibility check or supplement to the decoding attempts takes place.

The evaluation unit is preferably designed to execute only the one corresponding bit-logical operation with knowledge of the polarity of the codes and the polarity of destruction of the codes and to decode only the one common code resulting therefrom. In order to accelerate the evaluation, only the appropriate bit-logical operation can be carried out and evaluated. As an example, it is known that the codes are dark on a light background and that the main cause of destruction is in stray light, so light destruction should be repaired. Then the codes are represented by the binary "0" and the destructions by the binary "1". The appropriate bit-logical operation is then an OR of the negated binarized image data. In this way, as desired and necessary, the dark code areas are transmitted into the light destruction. Similarly, the appropriate bit-logic operation can be easily found for any combination of polarities of code and destruction.

The evaluation unit is preferably designed to execute only the two corresponding bit-logical operations with knowledge of the polarity of the codes or the polarity of destruction of the codes and to decode the respectively resulting common codes. So here is less information about the polarities. Therefore, an additional second bit-logic operation must be performed to cover both unknown polarity scenarios.

The evaluation unit is preferably designed to first attempt to decode the image data of the first code area and / or of the second code area separately and only if this fails to form a common code and to decode. The combination of redundant code areas and the subsequent decoding attempt of the resulting common code are thus executed only if previously a simple decoding attempt has failed, which uses only a single code area. This is advantageous if heavily destroyed codes remain the exception, because then the more complex process is rarely needed. Alternatively, the reading can take place immediately and directly via redundant combined code areas. This saves you the preliminary simple reading attempt, but the more complex reading over combined code areas always has to be done. Information is not lost compared to the simple read attempt when the four bit-logic operations are applied, since at least one of these operations correctly associates the code areas even if there are no major destructions.

The invention will be explained in more detail below with regard to further features and advantages by way of example with reference to embodiments and with reference to the accompanying drawings. The illustrations of the drawing show in:

1 a schematic sectional view of a camera-based code reader;

2 a schematic overview of a code reader on a conveyor belt with code-bearing objects in relative movement to the code reader;

3a a representation of a code with a destruction using the example of a data matrix code, the information blocks are shown purely schematichich;

3b a representation of a code according to 3a with destruction elsewhere in the code; and

3c a representation of a common code consisting of a combination of codes according to 3a and 3b arises and in which the destruction is compensated.

1 shows a schematic sectional view of a camera-based code reader 10 , The invention is based on the example of the camera-based code reader 10 Although it is also conceivable that image data is obtained with another light receiver, such as by scanning with a read beam.

The code reader 10 captures a recording area 12 in which are codes to read 14 can be located. The light from the recording area 12 is through a taking lens 16 received, in which only one lens shown 18 represents the recording optics. An image sensor 20 For example, a CCD or CMOS chip having a plurality of pixel elements arranged in a row or a matrix generates image data of the recording area 12 and gives these as raw image or input image to as a whole with reference numerals 22 marked evaluation unit on. For a better capture of the code 14 the code reader can be equipped with an active illumination, not shown.

The evaluation unit 22 is implemented on one or more digital devices, such as microprocessors, ASICs, FPGAs or the like, which are also wholly or partly outside the code reader 10 can be provided. The evaluation unit 22 includes a preprocessing or segmentation unit 24 , a binarization unit 26 , a combination unit 28 and a decoding unit 30 ,

In the segmentation unit 22 Image areas are identified as code areas in which codes to be read 14 could be present, for example, based on contrasts or variances. In addition, the image data can be edited, that is, for example, smoothed, brightness normalized or the like. The image data in the code areas are then in the Binarisierungseinheit 24 binarized, ie it becomes a black and white image from the gray value data of the image sensor 20 generated. In other embodiments, the binarization takes place before the segmentation, or it is waived and worked with grayscale or color image data.

The operation of the combination unit 28 will be related below 3 explained in detail. In a nutshell, image data from several code areas are combined here with mutually redundant codes in order to obtain a common code in which destructions in the individual underlying code areas balance each other out.

The common code then becomes a decoding unit 30 supplied to read the code content. This may be repeated with different common codes merged under different conditions to increase the likelihood of successful reading on partially corrupted image data.

At an exit 32 of the code reader 10 the read code information can be read out. The exit 32 or another, not shown interface can also be used to output raw image data or image data in different processing stages.

2 shows the code reader 10 in an important exemplary application, namely in a mounting on a reading tunnel or on a conveyor belt 34 which objects 36 as by the arrow 38 indicated by the field of view of the code reader 10 promotes. As throughout the description, like features have been given the same reference numerals. The objects 36 wear codes on their outer surfaces 14 that from the code reader 10 be read and evaluated. These codes 14 can from the code reader 10 only be read if they are mounted on the top or at least recognizable from above. Therefore, deviating from the illustration in 2 for reading an approximately laterally or down attached code 14a a plurality of code readers 10 be mounted from different directions, to allow a so-called Omnilesung from all directions.

Because of the combination unit 28 is the code reader 10 able to take advantage of code redundancy codes 14 which otherwise would not be readable due to artifacts in the code area or other partial destruction of the image data in the code area. This code redundancy may be present because of the same code 14 several times on the object 20 and thus exists in a recorded single image.

Alternatively, the code redundancy is used in a picture sequence in which the code 14 is recorded several times in succession. This happens among other things when reading in motion, for example in the in 2 presented application. In such a situation, the code redundancy is implicitly given without overhead. For example, is a specific area in the reading field of the code reader 10 outshone, no evaluable image data is obtained here. Due to the relative movement of the objects 36 However, this defect migrates across the code area, so that in different images of the image sequence the destruction of different parts of the code 14 meets. Thus, the destruction prevents the reading of the code 14 in every single frame. By exploiting the code redundancy in the image sequence, the code can 14 nevertheless be read.

Based on 3 now becomes the combination of code areas in the combining unit 28 described. As an example, the structure of a data matrix code is shown, although other 2D codes and 1D codes can also be read.

To actually improve the read rate, some conditions should be met. If some of these conditions are not fulfilled, the procedure will continue to apply, but the benefits may not be achieved in whole or in part, even under unfavorable conditions.

First, the code should 14 be present multiple times in a single frame or in a picture sequence, because without such redundant code data the basis for a combination is missing. The destruction should be locally disjoint, that is, they are in different places in the code areas to be combined. Subareas that are destroyed in all code areas can not be reconstructed by combination. After all, the destruction should be monochrome, thus clearly classifiable as light or dark. This is almost always the case with real disruptive effects. Otherwise, one could subdivide the code areas and first combine the subareas individually and then reassemble them so that the condition in the subareas is met.

3a shows the structure of a data matrix code 40 , whose codewords 42 illustrated by empty blocks. In a real data matrix code 40 would be the codewords 42 not empty but filled with the actual code information in the form of black squares at the positions corresponding to the code information. A destruction shown schematically as a black bar 44 hides part of several codewords 42 , It is assumed that the destruction 44 exceeds the error capacity, for example a Reed-Solomon correction in the decoding unit 30 , In practice, this still depends on what actual black and white information is underneath the destruction 44 is hidden or remains readable.

Unless the code 14 . 40 despite the destruction 44 remains readable, it does not require a combination of redundant code areas. Therefore, embodiments are conceivable in which the code is first tried 14 without reading combination. Only if this fails, redundant code areas are combined. On the other hand, it would be harmless to combine without exception, as undamaged codes 14 Of course, even after the combination remain legible.

3b shows another redundant image of the data matrix code 40 , The destruction 44 now concerns another subarea. Such a relative wandering of destruction 44 occurs in successively recorded images in an application such as in the 2 regularly on. But even if the data matrix codes 40 several times on an object 36 are appropriate and the code areas are from the same picture, it is at least extremely unlikely that a destruction 44 identical subregions meets.

The combination unit 30 now sets from the two data matrix codes 40 a common code 46 together, in 3c is shown and in which the destruction 44 eliminated or at least reduced. Conceivable for this are arbitrary methods, such as an averaging on the level of gray values.

In a preferred embodiment, the code areas are the data matrix codes 40 including the destruction 44 previously in the binarization unit 26 transformed into a pure black and white picture. Thus, there is a bit matrix M1 and a bit matrix M2 of the two code areas, which in 3a respectively 3b are shown. These two bit matrices are now combined with a bitlogical or boolean operation to create a new bitmatrix in which to destroy 44 missing information in the one matrix M1 are transferred from the other matrix M2. The information already contained identically in both code areas remains after the bit-logical operation, so that only the error locations are corrected.

Which bit-logical operation is the right one depends on the polarity of the code 40 and the polarity of destruction 44 from. Polarity refers to the brightness, so whether the code 40 or the destruction 44 is light or dark. In the in 3 example shown are both the data matrix code 40 as well as the destruction black, thus represented by a binary "0". The latter is of course just convention and therefore exemplary. The appropriate bit-logic operation for the example is an "OR". This leaves parts of the data matrix code 40 unaffected, which are contained in both redundant code areas. Under the destruction 44 In the bright intermediate areas of the code, a binary "1" hits a binary "0" of destruction 44 , so that the bright information of the code intermediate areas is just reconstructed by "OR" linking.

Depending on the polarity of the code 40 and the destruction 44 another bit-logical operation is required. Knowing the polarities, the required bit-logical operation can be determined in advance. If one knows only one polarity of the code 40 or destruction 44 , so two complementary bit-logic operations are required.

If both polarities are unknown, the following four bit-logic operations are performed: M1 AND M2, M1 OR M2,! M1 AND! M2,! M1 OR! M2. In this case, "!" Means a logical negation or inversion, ie an exchange of each binary "0" by the binary "1" and vice versa. This achieves independence from whether the code is bright on a dark background or vice versa, i. what polarity the code has.

One of the four resulting bit matrices is then, provided that the conditions mentioned above are given, always decodable, because the missing information of the two matrices on the monochrome destruction 44 was transferred to the combination matrix. The decoding unit 30 tries to decode each of the combination matrices therefore undertakes four read attempts in the general case and two or only a reading attempt with prior knowledge of the polarities. The further decoding attempts can be aborted once the code has been read once.

In a practical implementation, the evaluation unit may 22 memorize the respective bit matrices, if coded redundancies are to be used over a picture sequence. With code redundancy within a single frame, it is sufficient to remember the vertices of the detected code candidates and to form meaningful pairings, for example, based on similar geometries and surface contents. On the basis of the found meaningful pairings and the stored vertices, two bit matrices M1 and M2 can then be respectively generated, from which the combination matrices are calculated with the described bit-logical operations. The decoding attempts are then initiated with the combination matrices. Under certain circumstances, after a decoding process, two good readings may automatically be obtained on the original regions of M1 and M2.

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• EP 1096416 B1 [0007]

Claims (13)

Optoelectronic code reader ( 10 ) with a light receiving element ( 20 ) for the conversion of received light into image data and with an evaluation unit ( 22 ), which is adapted to code areas ( 14 . 40 ) in the image data and code information from the code areas ( 14 . 40 ), characterized in that the evaluation unit ( 22 ) is designed to store in the image data a first code area ( 14 . 40 ) and a second code area ( 14 . 40 ), the image data from the first code area ( 14 . 40 ) and the second code area ( 14 . 40 ) assuming that mutually redundant codes ( 40 ), to a common code ( 46 ) and then the common code ( 46 ) to decode the code information. Code reader ( 10 ) according to claim 1, which is a camera-based code reader ( 10 ), wherein the light receiving element ( 20 ) is a matrix or line image sensor. Code reader ( 10 ) according to claim 1 or 2, wherein the evaluation unit ( 22 ) is adapted to code areas ( 14 . 40 ) with two-dimensional codes ( 40 ) and two-dimensional codes ( 40 ) to decode. Code reader ( 10 ) according to one of the preceding claims, wherein the image data is an image with a plurality of code areas ( 14 . 40 ) exhibit. Code reader ( 10 ) according to one of claims 1 to 3, wherein the image data comprises a plurality of images with code regions ( 14 . 40 ), in particular a plurality of images of the same code-carrying object ( 36 ) in different stages of movement. Code reader ( 10 ) according to one of the preceding claims, wherein the evaluation unit ( 22 . 26 ) is adapted to the image data and in particular the code areas ( 14 . 40 ) to binarize. Code reader ( 10 ) according to claim 6, wherein the evaluation unit ( 22 ) is adapted to the code areas ( 14 . 40 ) to negate. Code reader ( 10 ) according to claim 6 or 7, wherein the evaluation unit ( 22 ) is adapted to code areas ( 14 . 40 ) or negated code areas by logical AND or by logical OR to a common code ( 46 ) to combine. Code reader ( 10 ) according to one of claims 6 to 8, wherein the evaluation unit ( 22 ) is adapted to the first code area ( 14 . 40 ) and the second code area ( 14 . 40 ) by means of at least one of four bit-logical operations: by logically ANDing the binarized image data of the first code area ( 14 . 40 ) and the second code area ( 14 . 40 ) - by logical ORing the binarized image data of the first code area ( 14 . 40 ) and the second code area ( 14 . 40 ) - by logically ANDing the negated binarized image data of the first code area ( 14 . 40 ) and the second code area ( 14 . 40 ) - by logical ORing the negated binarized image data of the first code area ( 14 . 40 ) and the second code area ( 14 . 40 ). Code reader ( 10 ) according to claim 9, wherein the evaluation unit ( 22 ) is designed to perform all four bit-logic operations and the respective resulting common codes ( 46 ) to decode. Code reader ( 10 ) according to claim 9, wherein the evaluation unit ( 22 ) is trained to know the polarity of the codes ( 40 ) and the polarity of destruction ( 44 ) of the codes ( 40 ) execute only the one corresponding bit-logical operation and only the one common code resulting therefrom ( 46 ) to decode. Code reader ( 10 ) according to claim 9, wherein the evaluation unit ( 22 ) is trained to know the polarity of the codes ( 40 ) or the polarity of destruction ( 44 ) of the codes ( 40 ) execute only the two corresponding bit-logical operations and the respective resulting common codes ( 46 ) to decode. Code reader ( 10 ) according to one of the preceding claims, wherein the evaluation unit ( 22 ) is adapted to first try to process the image data of the first code area ( 14 . 40 ) and / or the second code area ( 14 . 40 ) to decode for itself and only if this fails a common code ( 46 ) and to decode.

Images