JPH06266879A - Bar code detector - Google Patents

Abstract

PURPOSE:To detect a bar code in a frame memory by the same algorithm independently of the sort of the bar code and to detect the bar code even when its sort is not previously known. CONSTITUTION:A binary transformation part 16 divides picture data stored in the frame memory into blocks each of which consists of 8X8. picture elements and transforms respective blocks into binary data. A discrete cosine transformation (DCT) part 18 applies DCT to each block consisting of 8X8 picture elements and a coefficient memory 20 stores the transformed result (DCT coefficient). The binarization and DCT are applied to all blocks stored in the memory 14. A bar code detecting part 22 detects a bar code area from the DCT coefficient of each block stored in the memory 20.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a bar code detecting device for detecting a bar code from image data containing a bar code image.

[0002]

2. Description of the Related Art In recent years, POs in supermarkets, etc.
With the remarkable spread of the S (Point of Sales) system, barcodes have become ubiquitous. There are several tens of types of barcodes, which are roughly classified into one-dimensional barcodes and two-dimensional barcodes.

A one-dimensional bar code has a structure in which bars and spaces are alternately arranged in one line. In addition to JAN as shown in FIG. 3A, ITF, CODE128, CO
There are various code systems such as DE39.

However, since the amount of information that can be stored in these one-dimensional bar codes is as small as about a dozen or more characters, especially in recent years
A new barcode with a large amount of information called a dimensional barcode has been developed. For example, PDF417 as shown in FIG. 3B, DAT as shown in FIG.
ACODE, VERICOD as shown in (D) of FIG.
Various code systems such as E and Code16K have been proposed. These two-dimensional barcodes increase the amount of information by stacking one-dimensional barcodes and developing data in a matrix.

The decoding method of each bar code differs depending on the shape of the bar code. For example, in a one-dimensional bar code, generally, one of the lasers having a scanning pattern in a plurality of directions is configured to scan the entire bar code, or a line sensor is contacted to scan the entire bar code. In this method, the blank areas existing on both sides of the barcode are detected, the barcode data is determined, and the barcode data is decoded. Further, in recent years, as disclosed in, for example, Japanese Patent Application Laid-Open No. 2-23484, image processing such as contour detection is performed from image data including a barcode image to extract a barcode area, and a barcode is extracted. A method of determining a scan line to be traversed and decoding the scan line has also been proposed.

In the case of PDF417, the bar code is picked up by the area sensor and temporarily stored in the frame memory. Then, the data in the frame memory is read out in the horizontal and vertical directions to scan the rows, detect the start / stop codes 100 and 102 that are characteristic of the PDF417, determine the position of the bar code, and the bar code exists. Decoding is performed efficiently by finely scanning only the area. Alternatively, since the PDF417 has a structure in which one-dimensional barcodes are stacked, it can be decoded by swinging the laser in two dimensions and scanning each row.

In DATACODE and VERICODE, a bar code image is picked up by an area sensor, a frame memory is scanned, and a characteristic pattern of each code, that is, an L-shaped frame 104 and VERICODE of DATACODE.
The rectangular outer frame 106 of E is detected, the position of the bar code is determined, and the bar code is decoded.

As described above, in the system in which a bar code is imaged by the area sensor, the data is stored in the frame memory, and decoded, it is unknown at which position in the frame memory the bar code image exists. However, it is impossible to scan the frame memory in all directions, and the processing time is very long, which is not realistic. Therefore, a method of detecting the position of the bar code in the frame memory and scanning and decoding only the area where the bar code exists is generally proposed.

Further, for example, in a system in which an article on which a bar code is printed or pasted flows on a belt conveyor and an area sensor arranged at a predetermined position captures an image of the bar code and decodes the bar code, the decoding process generally takes time. Therefore, the key to speeding up is to perform the decoding process only when the barcode comes. Therefore, various methods have been proposed in which the presence of the barcode itself in the frame memory is detected and the decoding is started only when the barcode is present.

[0010]

However, in order to detect the presence or position of the bar code in the frame memory,
Pre-scanning for proper scanning in the frame memory is required, which is a simple and generally used method for coarse and full-screen scanning every few pixels in the horizontal and vertical directions. If the pre-scanning for bar code detection is improperly performed, it may take a long processing time or the bar code may not be detected.

Further, as a bar code detection method, in an algorithm for finding a characteristic pattern of each code from scan data, the detection algorithm is different for each code. Therefore, it is necessary to prepare a plurality of detection algorithms according to the kind of bar code to be decoded. There is.

Further, the above-mentioned algorithm has a problem that it is not known which pattern should be found unless it is known which bar code is about to be decoded.

The present invention has been made in view of the above points, and the same algorithm can be used to detect any type of bar code in the frame memory, and the bar code type can be determined in advance. It is an object of the present invention to provide a bar code detecting device that can detect even if it is not known.

[0014]

In order to achieve the above object, a bar code detecting apparatus of the present invention comprises a frame memory for storing multi-valued image data including a bar code image, and an image stored in the frame memory. D that performs discrete cosine transform (DCT) on each block of a predetermined size
It is characterized by comprising CT means, coefficient storage means for storing the DCT transform coefficient obtained by the DCT means, and barcode detection means for detecting a barcode from the transform coefficient stored in the coefficient storage means. It is a thing.

[0015]

According to the bar code detecting device of the present invention,
The DCT means subjects the image data stored in the frame memory to discrete cosine transform for each block of a predetermined size, and stores the DCT transform coefficient obtained by this DCT in the coefficient storage means. Then, the barcode detection means detects the barcode from the conversion coefficient stored in the coefficient storage means.

That is, the DCT is performed after binarizing the multi-valued data to eliminate the fluctuation of the pixel value in the black / white area.
And decompose into frequency components. Then, the barcode is detected by extracting the characteristic of the barcode area from the frequency component (DCT conversion coefficient).

[0017]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Before describing the embodiments of the present invention, first, the basic concept of the bar code detecting method of the present invention will be described. The principle is to detect the barcode area from the frequency component of the image data.

The DCT is a kind of high-efficiency coding that compresses the information amount of image data by dividing the image data into frequency components and performing quantization and coding for each component. This DC
In particular, T is a method most often used as a data compression technique for moving images and still images in recent years.

The DCT transform coefficient obtained by dividing the image data into a block of n pixels × n pixels and subjecting each block to DCT is obtained by determining what frequency component and how strong the data in the block is. Represents the squid. For example, if the data in the block is uniform, all AC components are zero and only DC components (offset) appear. That is, if the inside of the block is completely black or completely white, only DC components having different DCT transform coefficient magnitudes are obtained, as shown in FIG. 1B or 1C. On the other hand, if black-and-white data is mixed in the block, the DCT transform coefficient has a complicated value, as shown in FIG.

Therefore, since the bar code is composed of a black bar and a white space, when DCT is applied to the image data including this bar code, the DCT transform coefficient is converted into a black area, a white area, and a black and white mixed area. Can be classified. That is,
Since the AC component exists in the block in which the barcode, characters, etc. exist, it is possible to distinguish the background or the white area of the quiet zone of the barcode from the barcode, characters, etc. And since the bar code has a square shape and the bar and space have a length, the same coefficient is lined up,
Characters have various shapes, so that various frequency components appear in each block and no specific coefficient can be obtained, so that it is possible to distinguish the bar code from other characters. We think that the barcode area can be detected. An embodiment of the present invention according to the above idea will be described below with reference to the drawings. (First embodiment)

FIG. 1A is a block diagram of the bar code detecting apparatus 10 according to the first embodiment of the present invention. In the figure, reference numeral 12 is an input terminal for image data (multi-valued), and 14 is a frame memory for storing the input image data. Reference numeral 16 is a binary conversion unit that divides the image data stored in the frame memory 14 into n × n blocks and binarizes the data in the block with a certain threshold value, and 18 is the binarized n Dxn block data
The DCT unit 20 for CT is a coefficient memory for storing one screen of DCT transform coefficients output from the DCT unit 18. 22 is the DC stored in the conversion coefficient memory 20.
This is a bar code detection unit that detects a bar code area from the T conversion coefficient and reads out and supplies the image data of the bar code area from the frame memory 14 to the decoding unit 24.
The decoding unit 24 uses the data read from the frame memory 14 by the barcode detection unit 22 to
The barcode information is decoded, and the decoding result is transferred and output to a host device (not shown) or the like. Next, the operation of the first embodiment will be described.

That is, image data including a bar code image picked up by a CCD (not shown) or the like is input from the input terminal 12 and stored in the frame memory 14 as shown in FIG. The image data stored in the frame memory 14 is divided into, for example, blocks of 8 × 8 pixels of the frame memory 14 by the binary conversion unit 16 and converted into binary data. Since the input image is multi-valued data, the value is subtly fluctuated for each pixel even if it is a black bar or white space. Therefore, if DCT is applied as it is,
Since the AC component appears even in the flock that can be regarded as black, the binary conversion unit 16 performs binarization to eliminate the fluctuation of the data value. Next, in the DCT unit 18, 8 × 8
DCT of a block of pixels is performed, and the conversion result (DCT
The conversion coefficient) is stored in the coefficient memory 20. This binarization,
The DCT is performed on all blocks, and the barcode detection unit 22 detects the barcode area from the DCT transform coefficient of each block. For example, a black block is extracted from the DC component of each block, the neighboring blocks are grouped, an area forming a substantially quadrangle is detected, and this is determined as a barcode area. Then, since the block including the barcode, that is, the position of the barcode is known, the data corresponding to the area is read from the frame memory 14, and the decoding unit 24 decodes the barcode information by the conventional decoding process. , Transfer the decoding result.
In this way, the image data including the barcode image is decomposed into frequency components, and the features of the frequency components are extracted, whereby the barcode in the image can be detected. (Second Embodiment) As a modification of the first embodiment, there is a method of detecting a characteristic pattern of various bar codes in the processing of the bar code detecting section 22.

Specifically, if blocks having similar frequency components are arranged in a certain direction and a number of columns appear in parallel, it is determined that a one-dimensional bar code exists. Also, in the square area that is seen as the barcode area,
If there is a block group representing a large black bar, PDF4
Judge as 17. Separately, if there is a block group representing a black L-shape or a square black frame, DATACODE or VE
It is determined to be RICODE, the data in the frame memory 14 is read by the decoding method according to each code, and the decoding unit 24 performs the decoding process.

As described above, by decomposing the image data containing the bar code image into frequency components, the characteristics of the bar code having various shapes can be expressed by the frequency components. Therefore, by capturing the characteristics, the bar code Can be identified. (Third embodiment)

As another embodiment, with the same configuration as the first embodiment, image data is repeatedly picked up by an image sensor such as a CCD, and only when an image is picked up, the bar code detection and decoding processing is started. Think of the system.

When nothing is imaged, the screen is entirely white or black, so that the AC component of the DCT transform coefficient does not appear anywhere. Therefore, in the process of the bar code detecting unit 22 in the first embodiment, if there is no AC component, it is determined that no image is picked up, and the process proceeds to the next image capture. On the other hand, if the AC component exists,
It is determined that an image has been picked up, and the process moves to the barcode area detection and decoding processing.

As described above, by analyzing the frequency component of the imaged image data, it can be detected that an object (bar code) has entered the screen, and the automatic reading of the bar code can be started. (Fourth Embodiment) Another embodiment will be described with reference to the block diagram of the bar code detecting device 30 in FIG.

In the figure, reference numeral 12 is an input terminal for image data (multivalued), and 14 is a frame memory for storing the input image data. Reference numeral 16 denotes a binary conversion unit that divides the input image data stored in the frame memory 14 into m × m pixel blocks of the frame memory and binarizes the data in each block with a certain threshold value.
Is the size of m that can fill the space of the barcode.
It is an expansion processing unit that expands black with a size of × m. Reference numeral 18 denotes n × n data obtained by expanding the black in the expansion processing unit 32.
(N <= m) DCT unit for DCT in block, 20 is D
This is a coefficient memory that stores CT conversion coefficients for one screen. Two
Reference numeral 2 is a bar code detecting section for detecting a bar code area from the DCT transform coefficient, and 24 is a decoding section for decoding bar code information from the data read from the frame memory 14. Next, the operation of the fourth embodiment will be described.

The image data input from the input terminal 12 is stored in the frame memory 14. The image data stored in the frame memory 14 is converted into binary data by the binary conversion unit 16. Next, the expansion processing unit 32 performs black expansion processing on the m × m size scanning window. This is a general method, and the black pixel is expanded by the size of the scanning window by taking the logical sum of the data in the scanning window. Then, in the barcode area, the space is filled with black and becomes one large black square. The DCT unit 18 converts the image data including the black bar code image into n
The DCT is applied to the × n block, the size of which is smaller than or equal to the scanning window size m × m at the time of expansion processing. Then, in the DCT transform coefficient, a block having a square AC component appears at a position corresponding to the peripheral portion of the barcode area, and the inside thereof becomes a block group including only the DC component representing a black block. Therefore, in the barcode detection unit 22,
By extracting the square black block group, the barcode area can be detected.

As described above, since the bar code area is black-expanded to form a square black area, it is possible to give characteristics by the generation of the frequency component, so that the bar code detection processing can be easily performed. (Fifth Embodiment) In the fourth embodiment, after the expansion processing, data thinning is performed and then DCT is performed. Then, since the number of blocks is reduced, the amount of calculation is reduced and higher speed processing can be performed.

In the first to fifth embodiments described above,
The DCT transform coefficients stored in the coefficient memory 20 do not necessarily have to be all coefficients, and may be only DC components or only DC components and low-frequency components of AC.

[0032]

As described in detail above, according to the present invention,
Decompose the image data including the barcode image into frequency components,
Since barcode detection is performed by extracting the characteristics of frequency components, the same algorithm can be used to detect barcodes in the frame memory regardless of the type, and the barcode type is known in advance. It is possible to provide a bar code detection device that can perform detection without the bar code. Further, the binarization by the binary conversion unit 16 can be omitted.

[Brief description of drawings]

FIG. 1A is a block configuration diagram of a bar code detecting apparatus according to a first embodiment of the present invention, and FIGS. 1B to 1D are DCs obtained by discrete cosine transform (DCT).
It is a figure for demonstrating a T conversion coefficient.

FIG. 2A is a schematic diagram of image data including a barcode image stored in a frame memory, and FIG. 2B is a block configuration diagram of a barcode detection device according to a fourth embodiment of the present invention.

FIGS. 3A to 3D are diagrams showing barcodes, respectively.

[Explanation of symbols]

10, 30 ... Bar code detecting device, 12 ... Input terminal, 1
4 ... Frame memory, 16 ... Binary transform unit, 18 ... Discrete cosine transform (DCT) unit, 20 ... Coefficient memory, 22 ...
Bar code detection unit, 24 ... Decoding unit, 32 ... Expansion processing unit.

Claims (1)

[Claims] 1. A frame memory for storing multi-valued image data including a bar code image, and a discrete cosine transform means for performing discrete cosine transform of the image data stored in the frame memory for each block of a predetermined size. A coefficient storage unit that stores the conversion coefficient obtained by the discrete cosine conversion unit; and a bar code detection unit that detects a bar code from the conversion coefficient stored in the coefficient storage unit. Bar code detection device.