JPH09185671A - Device and method for recognizing bar code - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a bar code recognizing device with which a bar code label stuck on a moving object can be recognized with high accuracy even when the bar code label is passed in any unspecified direction. SOLUTION: Images fetched from a belt conveyer 4 for moving a moving object 5, to which a bar code label 6 is stuck, and plural multicamera units 3 installed at the upper part of belt conveyer 4 are sent to a bar code recognizing device 1. At the bar code recognizing device 1, the input images are filtered and the bar code is extracted from the images fetched from the multicamera units 3. Then, the inclination of extracted bar code is detected, corresponding to this inclination, the feature amount of bar code is extracted and based on this extracted feature amount, the bar code is recognized.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a bar code recognition apparatus by image processing, and particularly to a recognition method and apparatus suitable for reading a bar code attached to a moving body passing through a wide field of view in a specific direction. Regarding

[0002]

2. Description of the Related Art Recently, in order to easily control the quality of products, a barcode is attached to the product, this is captured as an image, and the barcode attached to the product is read to identify the type of the product. There is a way to recognize and manage the product.

A technique for detecting a code attached to a moving body carried by a carrying device is, for example, Japanese Patent Laid-Open No. 58-19.
It is described in Japanese Patent No. 5275. This technology relates to products that are delivered, and when the code attached to the product is captured by a camera to recognize the code, it can be read even if the code angle is not constant. Is. It forms the code in groups of parallel stripes detected at an angle of 90 ° to 45 ° with respect to the stripe direction, with one alignment stripe attached to the group. At the time of detection, the code is evaluated by locating the position of the code based on this positioning stripe.

[0004]

However, forming the stripes for determining the position of each code as described above increases the size of the entire code and the size of the entire label. Also, when a code without this stripe comes in, the code cannot be read, and the versatility is lacking. Further, in the above-mentioned conventional technique, since the code is taken in by one image input device, when the product is carried on a conveyor, for example, there are cases where the product cannot be taken in successfully.

Further, when recognizing the bar code label captured from the camera in this way, noise other than the bar code appears in the image. In such a case, if the noise and the barcode are not separated, the noise portion may be recognized as a barcode, resulting in erroneous recognition.

The present invention has been made in view of the above drawbacks. Even when a bar code label attached to an object moving at high speed passes in a wide field of view in an unspecified direction,
An object of the present invention is to provide a barcode recognition device that reliably extracts barcode labels in a narrow installation environment and recognizes barcodes with high accuracy.

[0007]

[Means for Solving the Problems] The above object is to provide a transport device for moving an object to be measured with a bar code attached thereto, a plurality of image input devices installed above the transport device, and the image input device. A bar code recognition device comprising: an image processing device for processing the captured image to read the barcode attached to the measurement object; and filtering the images captured from the plurality of image input devices. A bar code extracting section for extracting the bar code from the processed and filtered image and the image captured from the image input device; and a bar code scanning processing section for extracting the feature amount of the extracted bar code. This can be achieved by including a processing device having a barcode recognition unit that recognizes the barcode based on the extracted feature amount. .

[0008] Further, the above object is to capture an image of an object to be measured to which a bar code is attached from an image device, process the captured image, and display the bar code attached to the object to be measured. In the reading barcode recognition method, an image captured from the plurality of image input devices is filtered, the barcode is extracted from the filtered image and the image captured from the image input device, and the extracted barcode is extracted. This can be achieved by detecting the inclination of the barcode, extracting the characteristic amount of the barcode according to the inclination, and recognizing the barcode based on the extracted characteristic amount.

[0009]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 shows the configuration of the entire system of a bar code recognition apparatus by image processing.

In FIG. 1, the bar code recognition device 1 is
The barcode label detection processing is always performed on the image input from the multi-camera unit 3. In the normal state, the barcode image cannot be detected because the input image is dark. Therefore, the strobe 2 is provided to detect the barcode label, and the input image is brightened to perform the barcode label detection processing. Specifically, the moving body 5 to which the barcode label 6 with the barcode printed is attached is conveyed from the belt conveyor 4. When the moving body 5 passes through the optoelectronic switch 7, the signal from the optoelectronic switch 7 is sent to the barcode recognition device 1.
As a result, the barcode recognition device 1 determines that the moving body 5 as the recognition target object has entered the recognition area where recognition is possible. The signal transmitted from the photoelectric switch 7 to the barcode recognition device 1 causes the barcode recognition device 1 to move to the flash unit 2.
A strobe ON signal is transmitted to a synchronizing circuit section (not shown) in the multi-camera unit 3 in order to cause the flash. The synchronizing circuit section in the multi-camera unit 3 transmits a synchronizing signal to the strobe 2, and the strobe 2 emits light in accordance with this synchronizing signal. When the strobe 2 emits light, it becomes brighter, so that the barcode printed on the barcode label 6 can be recognized by the barcode recognition device 1. When the moving body 5 to which the barcode label 6 is attached passes under the multi-camera unit 3 in the state in which the barcode label 6 can be detected as described above, the barcode recognition device 1 causes the barcode label 6 to be detected. Is detected and recognition processing is executed. A plurality of cameras and a synchronization circuit are arranged in the multi-camera unit 3, and each camera captures an image in accordance with a synchronization signal from the synchronization circuit and transmits the captured image to the barcode recognition device 1. When the moving body 5 passes in front of the optoelectronic switch 8, the optoelectronic switch 8 sends an interrupt signal to the barcode recognition device 1. As a result, the barcode recognition device 1 recognizes that the moving body 5, which is the recognition target, has left the recognition area. The bar code recognition device 1 transmits a strobe OFF signal to a synchronizing circuit unit in the multi-camera unit 3 in order to extinguish the strobe 2 by an interrupt signal from the optoelectronic switch 8. As a result, the synchronizing circuit section in the multi-camera unit 3 stops the synchronizing signal transmitted to the strobe 2. When the strobe 2 is extinguished, the bar code recognition device 1 cannot detect the bar code label 6 because the input image becomes dark. When the barcode recognition device 1 receives the interrupt signal from the optoelectronic switch 8 during the detection process of the barcode label 6, the barcode recognition device 1 recognizes that the barcode label 6 is not attached to the moving body 5.

In this way, the light emission of the strobe 2 is controlled by the photoelectronic switch 7 and the photoelectronic switch 8, so that the life of the strobe 2 is extended and the strobe 2 can be used for a long time.

FIG. 2 shows the hardware configuration of the barcode recognition apparatus 1.

In FIG. 2, the hard disk 201 stores environment files necessary for barcode recognition. To store the data in this environment file, the operator refers to the display 205 to display the mouse 206 and the keyboard 20.
You can enter from 7 freely. The bar code recognition device 1 causes the processing device 203 to execute a bar code recognition process when the power is turned on. The processing device 203 includes a barcode extraction unit 1401, a barcode label extraction unit 1402, a barcode inclination detection unit 1403, and a line scan processing unit 1.
404, barcode decoding processing 1405, counter unit 14
06 and a memory unit 1407. The contents of the processing of each unit will be described later. The I / O controller 204 controls the input / output of man-machine type I / O of the display 205, mouse 206, and keyboard 207. The DI / DO board 216 inputs interrupt signals from the optoelectronic switches 7 and 8 and outputs a strobe ON / OFF signal to the multi-camera unit 3. The image processing boards 213, 214, and 215 perform image processing on image signals sent from two cameras, respectively. For example, in this embodiment, the image processing board 213 responds to the image signals from the cameras 218 and 219 while the image processing board 214 controls the camera 22.
The image processing board 215 performs image processing on the image signals from the camera 222 and the camera 221 respectively. Each of the image processing boards 213, 214, and 215 includes an image processing processor 211 that controls image processing, an image memory 209 that stores image data, and a camera switch 208 that is connected via a video interface 212.

FIG. 3 shows various moving bodies 302, 303, 304 carried by the belt conveyor 4 and the moving body 30.
Bar code label 6 attached to 2,303,304,
The structural example of the multi-camera unit 3 which makes a conveyor width the visual field range is shown.

In FIG. 3, a bar code label 6 is printed with a bar code, and various moving bodies 302, 30 are shown.
It is attached to 3,304. Barcode label 6
A camera field of view 35 indicates a plurality of camera field of view areas installed in the multi-camera unit 3 for recognizing the barcode printed on the. In this embodiment, a case where six cameras are provided will be described as an example. When the field width of one camera is 133 mm, the field width of six cameras is 598 mm. That is, the overlapping width between the cameras is 40 mm, and the width 36 of the target barcode label 6 is 36 mm.
It is larger than mm. By overlapping the visual field areas of the cameras in this manner, the barcode label 6 that has entered the overlapping portion can be reliably captured as an image by either camera. The moving bodies 302, 303, 304 actually moved by the belt conveyor 4 are oriented in arbitrary directions, and the barcode labels 6 attached to the movable bodies 302, 303, 304 are also oriented in arbitrary directions. ing. The motor 36 is for operating the belt conveyor 4.

Next, the processing for recognizing a barcode by the processing device 203 in the barcode recognition apparatus 1 will be described.

FIG. 4 shows the flow of the entire bar code recognition process.

In block 401, the input image captured from the multi-camera unit 3 is captured in the image memory 209 in the image processing board. In block 402, a striped pattern extraction image 410 is generated by extracting a black striped pattern, which is a characteristic of the barcode, from the input image 409 captured in the image memory 209 by the processing by the barcode extraction unit 1401 of the processing device 203. . In block 403, the striped pattern extraction image 410 extracted by the processing of the barcode label extraction unit 1402 of the processing device 203 is subjected to expansion processing and reduction processing to generate a barcode label extraction image 411 in which the area of the barcode label 6 is extracted. , Calculate the coordinates of the start and end points of the area. A block 404 detects the inclination of the barcode so that the barcode can be recognized even when the barcode label 6 is detected in an unspecified direction by the processing of the barcode inclination detection unit 1403 of the processing device 203.
In block 405, the line scan processing unit 1404 of the processing device 203 extracts the luminance, which is the pixel information of the barcode portion, according to the inclination of the barcode, and the extracted pixel information is used as a histogram to generate a line scan image. Blocks 406 to 408 are processing devices 20.
Block 4 in the processing of the barcode decoding processing unit 1405 of 3
07 determines whether or not the decoding of the bar code label has been repeated the number of times specified in advance in the environment file. If the number of times specified has not been specified, the processing of block 405 is performed again.

On the other hand, if the decoding is repeated the specified number of times, it is confirmed in block 408 whether or not the decoding result is correct based on the decoding result of the number of times specified in the environment file. Here, the method of determining whether or not the decryption result is correct is to judge that the decryption result is correct if the same decryption result is more than the majority of the decryption times out of the results obtained by performing the decryption a plurality of times, and if it is not, it is an error. Judge that there is.

FIG. 5 shows the details of the processing of the barcode extraction unit 1401. Only the black striped pattern which is the characteristic of the barcode is used, and the maximum value filter, the minimum value filter and the image difference of the image processing command are used. 3 shows a flow of a process of extracting by doing.

Here, the maximum value filter and the minimum value filter will be described. The maximum value filter selects one pixel from the target image, selects m × n pixels in the vicinity of this pixel, obtains the maximum value of the brightness in the m × n pixels, and selects the brightness value. The pixel value. Then, this processing is performed for all pixels. In addition, the minimum value filter selects one pixel from the target image, selects m × n pixels in the vicinity of this pixel, finds the smallest value of the brightness in the m × n pixels, and selects that value. The value of the selected pixel. Then, this processing is performed for all pixels.

The processing of the input image 509 will be specifically described as an example. Here, the input image 509 is the barcode (a)
And the image other than the barcode includes noise (b) larger than the barcode and noise (c) smaller than the barcode.

First, in block 501, the maximum value filter execution counter provided in the counter unit 1406 of the processing device 203 for counting the number of times the maximum value filter has been executed is cleared. In block 502, image processing is performed on the input image 509 by the maximum value filter of the image processing command. Block 503 increments the maximum filter enforcement counter. The block 504 determines whether or not the maximum value filter implementation counter has been implemented the number of times specified in advance in the environment file, and if it has not been executed the number of times specified, the process is performed again from the block 502.
In this way, when the processing by the maximum value filter is repeatedly performed on the input image, the barcode (a) and the noise (c) disappear, and an image 511 in which the noise (b) remains is obtained. However, since the thickness of the barcode varies depending on the size of the field of view of the camera, the number of times the maximum value filter for eliminating the striped pattern, which is a characteristic of the barcode, is also varied. Therefore, this problem is solved by registering the number of times until the black bar disappears by performing the maximum value filter in the environment file in advance.

When the maximum value filter process has been performed the specified number of times in block 504, block 505 and subsequent steps are executed. Blocks 505, 506, and 507 perform the minimum value filter for the number of times the maximum value filter has been executed. By performing processing with this minimum value filter, an image 511 is obtained in which the noise (b) of the image 511 having only the noise (b) is returned to the same original size as the input image 509. Then, in block 508, the image 512 resulting from the minimum value filtering and the input image 509
Take the difference with This removes the noise (b),
An image 513 in which the black bar portion of the barcode (a) and the noise (c) are highlighted is generated.

Thus, the noise (b) larger than the bar code (a) can be eliminated by combining the maximum value filter and the minimum value filter.

However, the noise (c) smaller than the bar code (a) is erased before the bar code (a) even if the maximum value filter is performed.
It cannot be processed in this way. Therefore,
FIG. 6 shows a process for eliminating noise smaller than that of the barcode (a).

FIG. 6 shows the details of the bar code label extraction unit 1402 of the processing device 203. The bar code (a) is shown in FIG.
7 shows a flow of processing for extracting a portion of the barcode label 6 from which a smaller noise (b) is removed from the input image.

In FIG. 6, blocks 601, 602 and
In 603, in order to perform the binarization processing of the image processing command on the image 513 extracted in FIG. 5, the maximum value and the minimum value of the luminance of the pixel are obtained and the binarization threshold value is calculated. Block 604 uses the binarization threshold to extract image 513
To a binary image 714 shown in FIG. Block 6
05, 606, 607, and 608 repeat the expansion process of the image processing command for the binarized image 714 a predetermined number of times. By executing the expansion process in this manner, the white area is expanded, and finally an image 715 in which the barcode is a continuous white area is generated.
A block 609 assigns a label to each white area by performing labeling processing of an image processing command on the white area in the shaped binarized image, and obtains an area of the white area to which each label is assigned. Block 610 obtains the label number of the maximum area from the white area to which the label is assigned. Block 611 erases all labels except the largest area label. That is, here, the image is blackened. As a result, the barcode (a) portion is extracted, and the image 717 in which the noise (c) is erased can be generated. Block 612 uses the x-axis projection of the image processing command to determine the x-coordinate (x
e, xs) is calculated, and the memory unit 140 of the processing device 203 is calculated.
Store in 7. Block 613 uses the y-axis projection of the image processing command to determine the y-coordinate of the label with the largest area.
(yes, ys) is calculated, and the memory unit 14 of the processing device 203 is calculated.
It is stored in 07. If the bar code is tilted, the x coordinate (xe, xs) and the y coordinate (y
e, ys). Thereby, the area coordinates of the barcode label in the input image can be calculated.

FIG. 8 shows the details of the barcode inclination detecting section 1403 of the processing device 203, and shows the flow of the processing for obtaining the inclination of the barcode label and extracting the pixel information.

In FIG. 8, block 801 is a starting point (xs, y) obtained by maximum area extraction in order to obtain the inclination.
s) and the end point (xe, ye) are read from the memory unit 1407 and a white area having the maximum area is enclosed (hereinafter, the enclosed area is referred to as a window). Block 802 is 45 °
In order to determine whether or not the label is inclined as described above, the horizontal side (| xe-xs |) and the vertical side (| ye-y) of the window frame are determined.
s |) are compared. If the horizontal length is longer than the vertical length, blocks 803 and 804 are executed. If the vertical side is large, blocks 807 and 808 are executed. A block 803 obtains the difference between the two representative points at the upper side of the window frame and the upper part of the white area, and the change amount of the difference (u2-u1).
Find the upper slope of the white area from. Block 809
Calculates the difference between the two representative points on the lower side of the window frame and the lower part of the white area, and calculates the inclination of the lower part of the white area from the change amount (d2-d1) of the difference. In blocks 807 and 808, the same processing as in blocks 803 and 804 is performed for the left side and the right side, and the inclinations of the right side and the left side are obtained. The block 805 is a block 803, 804 or a block 808, 809.
The average slope is obtained from the slopes of the upper side (left side) and the lower side (right side) obtained in.

Average slope = (upper (left) slope + lower (right side) slope) / 2 A block 806 determines a tangent when the determined average slope passes through the center of the window. Accordingly, the start point and the end point of the line scan process for extracting the pixel information can be obtained as follows, and the memory unit 14 of the processing device 203 can be obtained.
07.

Starting point X coordinate Lxs = xs Starting point Y coordinate Lys = Slope of average × xs + tangent side End point X coordinate Lxe = xe End point Y coordinate Lye = Slope of average × xe + tangent side As shown in FIG. The barcode is scanned to obtain the luminance thereof, and the histogram shown in FIG. 12 is created. However, when the barcode is inclined, the luminance may not be accurately obtained. FIG. 9B is an enlarged view of a part of the barcode. In this case, if the scan is performed in β, the pixels of the barcode can be captured, but in the case of scanning like α, the barcode is scanned. I cannot catch the pixels of the code. Therefore, in order to prevent such a case, the barcode image is displayed at 0 to 45 degrees as shown in FIG. 9C and 45 to 90 degrees as shown in FIG. 9D.
Complement in stages. The black part is the pixel showing the barcode, and the white part is the complemented part.

The scan processing unit 1 of the processing device 203
FIG. 10 shows a processing flow of the complementary method performed in 404.

First, block 1001 clears the array number used when data is stored in the pixel data storage array and the distance data storage array in the memory unit 1407 of the processing device 203. Block 902 determines if the barcode label tilt is within 45 degrees. If the angle is within 45 degrees, the process of sequentially changing the x-axis and fetching pixel information is performed according to the flow of blocks 1003 to 1015. 45
If it is not within the degree, a process of sequentially changing the y-axis and fetching pixel information is performed in blocks 1016 to 1028.
Follow the flow of. Block 1003 is the memory unit 1
The x coordinate of the start point of the loop counter (hereinafter x1) at 407 is set. A block 1004 initializes the y-coordinate of the starting point as a line feed determination element necessary for determining whether or not the y-axis coordinate has changed. The block 1005 calculates the y coordinate (y1) for fetching the pixel information at the position indicated by the x coordinate of the loop counter. Block 1006 is
It is determined whether the calculated y coordinate is the same as the line feed determination element. If they are the same, the processing of block 1011 and thereafter is executed, and if they are different, the complementary processing of blocks 1007 to 1010 is executed and then the processing of block 1011 and thereafter is executed.
A block 1007 performs a complementary process because the y-axis coordinate has changed. The complementary method takes in the brightness of the pixel indicated by the x coordinate immediately before the coordinate indicated by the loop counter and the y coordinate calculated in block 1005. Also, x indicated by the loop counter
The brightness of the pixel indicated by the coordinate and the y coordinate indicated by the line feed determination element is fetched. The average of the brightness of the two captured pixels is calculated and set in the image data storage array.

Image data storage array = ((x1-1, y
1) Pixel brightness + (x1, line break determination element) pixel brightness) / 2 The block 1008 complements the distance data from the start point as in the following equation and sets it in the distance data storage array.

Distance data storage array = (x1-start point x coordinate-0.5) / (cos (tilt angle)) The block 1009 increments the array number.
A block 1010 sets y1 in the line feed determination element and updates it. The block 1011 fetches the luminance of the pixel at the position indicated by the loop counter and y1 and sets it in the pixel data storage array. A block 1012 complements the distance data from the start point as in the following equation and sets the distance data in the storage array.

Distance data storage array = (x1-start point x coordinate) / (cos (tilt angle) The block 1013 increments the array number.
Block 1014 determines whether the loop counter has changed to the x coordinate position of the end point. If it has changed, the processing is terminated, and if it has not changed, block 1015.
Increments the loop counter at block 10
Execute after 05.

In the processing from block 1016 to block 1028, the y-axis coordinate is sequentially changed to block 1003.
The processing similar to the processing up to block 1015 is performed. As a distance data complement method, a flag array corresponding to the image data storage array of the complemented pixels may be provided, and the distance data may be complemented by the presence or absence of the flag after the pixel data is captured.

FIG. 11 shows a flow of processing for extracting the characteristic amount of the barcode from the read pixel information and converting it to a character.

In FIG. 11, a block 1111 finds a portion required for bar code decoding as a search range from the line-scanned pixel information. The method is to find the maximum and minimum values of luminance from the image data storage array that stores the pixel information, and detect the start and end position by adding the value obtained by adding 1/3 of the maximum and minimum values to the minimum value. Calculate as a value.
Next, the image data storage array is sequentially incremented from 0 to search for a brightness equal to or lower than the start end position detection value, and the array number storing the brightness value searched first is set as the start position of the search range. Further, the image data storage array is decremented sequentially from the end to search for a luminance value equal to or lower than the start / end position detection value, and the array number storing the first searched luminance value is obtained as the end position of the search range. Block 1102 calculates the position of the thick black bar. The calculation method will be described with reference to FIG. First, let the minimum value used in the search range calculation be the threshold value of the calculated value of the thick black bar. To calculate the thick black bar, the threshold value is gradually increased with respect to the luminance value within the search range shown between the star position and the end position, and the number of pieces of pixel information showing luminance below the threshold value is searched for. , Repeatedly updating the threshold value until the retrieved number becomes (recognition character specified in environment file + 2) × 2. The example of FIG. 12 is an example when the number of recognized characters is seven. When the luminance below the threshold value is searched, the distance data of the pixel is fetched from the distance data storage array and stored in the black thick bar position array. Block 1103 calculates the position of the white bar. The calculation method will be described with reference to FIG. First, the maximum value used in the search range calculation is used as the white bar calculation threshold value. The calculation of the thick white bar is performed by searching for the number of pieces of pixel information indicating the brightness equal to or higher than the threshold value while gradually decreasing the threshold value with respect to the brightness value within the search range shown between the start position and the end position. Iterate while updating the threshold value until the searched number reaches (the number of recognized characters specified in the environment file + 1). The example of FIG. 12 is an example when the number of recognized characters is seven. When the luminance equal to or higher than the threshold value is searched, the distance data of the pixel is fetched from the distance data storage array and stored in the thick bar position array. Block 11
04 calculates the position divided into barcode recognition character units. In the method of calculating the separated positions, the width of the black bar is calculated from the first and last values of the thick black bar position array calculated by first detecting the black bar position. The ratio of thin bar to thick bar is 1:
In the case of the relationship of 2.5, the width of the thin bar: the interval of one character has the relationship of 1: 14.5 (including the width of the character gap). From this relationship, the width of one character is calculated, the entire character delimiter position is calculated, and stored in the delimiter position array. A block 1005 calculates the characteristic amount of the end character “*”. The feature amount indicates the distance from the division position to the first thick black bar, the distance to the second thick black bar, and the distance to the thick white bar. The characteristic amount of the end character can be obtained by the following formula.

First thick black bar = (value before the last of the thick black bar position array)-(last of the delimiter position array) Second thick black bar = (the last value of the thick black bar position array)-
(End of delimiter position array) White bar = (End of white bar position array)-(End of delimiter position array) In block 1106, the feature amount calculated in block 1105 is the end character "*" feature. Quantity (Fig. 13
Indicates the list of bar code feature values) and the allowable value specified in the environment file. If they match, block 1110 and subsequent blocks are executed. If they do not match, blocks 1107 and 11
Reference numerals 08 and 1109 are processing when the bar code label is captured in a state of being rotated by 180 degrees with respect to the scan direction of the line scan processing, and the values of the black thick bar position array, the white thick bar position array, and the delimiter position array are recalculated. . Block 110
The recalculation method of 7, 1108 is performed for all data of the array as follows, and then rearranged in ascending order of the values.

Black bar position array = (last value of distance data storage array)-(value of black bar position array) White bar position array = (last value of distance data storage array)
-(Value of Value Array of White Bar) The block 1109 performs the same calculation as the block 1104 again to obtain the value of the delimiter position array. Block 111
In the case of 0, the feature amount is obtained in the character delimiter unit. The calculation method is the same as in block 1105, where the nth recognized character (n is 0
To the number of recognized characters-1) is obtained as follows.

First thick black bar = ((2 × n + 2) th value of the thick black bar position array)-(nth position of delimiter position array) Second thick black bar = ((2 × of the thick black bar position array) n + 3)
Value 0)-(nth position of break position array) white bar = (nth value of white position bar position array)-(nth value of break position array) Block 1101 is within an allowable value for the calculated feature amount. The character having the matching feature amount is searched from the feature amount management table. (The contents of the feature amount management table are shown in FIG. 13.)
Block 1112 determines whether or not the number of recognized characters specified in the environment file has been read. If the number of characters has been decomposed and read, the process is terminated, and if not, the block 1110 and subsequent steps are executed to decode the next character.

[0045]

According to the present invention, it is possible to inexpensively provide a device that can recognize a barcode attached to a moving body that moves in an unspecified direction on an existing production line or inspection line.

[Brief description of the drawings]

FIG. 1 shows the overall configuration of a barcode recognition system by image processing.

FIG. 2 shows a hardware configuration of a barcode recognition device.

FIG. 3 is a view showing a field of view of a multi-camera installed on a belt conveyor and a moving body to be carried.

FIG. 4 shows an overall flow of a barcode recognition procedure.

FIG. 5 shows a flow of a process of extracting a striped pattern which is a feature amount of a barcode.

FIG. 6 shows a processing flow for extracting a barcode label portion.

FIG. 7 shows an image when a barcode label portion is extracted.

FIG. 8 shows a processing flow for calculating the inclination of a barcode label.

FIG. 9 shows a method of complementing line scan processing.

FIG. 10 shows a flow of a line scan process for reading pixel information of a bar code portion.

FIG. 11 shows a processing flow for decoding a barcode.

FIG. 12 shows an example of black line extraction and white line extraction.

FIG. 13 shows a list of bar code feature amounts.

FIG. 14 shows a configuration of a processing device.

[Explanation of symbols]

1 ... Bar code recognition device, 2 ... Strobe, 3 ... Multi-camera unit, 4 ... Belt conveyor, 5 ... Moving body, 6 ...
Bar code label, 7, 8 ... Photoelectric switch, 201 ...
Hard disk, 203 ... Processing device, 213, 214,
215 ... Image processing board, 1401 ... Bar code extraction section, 1402 ... Bar code label extraction section, 1403 ... Bar code inclination detection section, 1404 ... Line scan processing section, 1405 ... Bar code decoding processing section, 1406 ... Counter section, 1407 … Memory section.

 ─────────────────────────────────────────────────── ─── Continued Front Page (72) Inventor Yoshiki Kobayashi 7-1, 1-1 Omika-cho, Hitachi-shi, Ibaraki Hitachi Ltd. Hitachi Research Laboratory (72) Inventor Tadaaki Kitamura 7-chome, Omika-cho, Hitachi-shi, Ibaraki No. 1 Hitachi Ltd. Hitachi Research Laboratory (72) Inventor Minoru Hisashima 52-1 Omika-cho, Hitachi City, Ibaraki Prefecture Incorporated Hitachi Ltd. Omika Factory (72) Inventor Takumi Ebisawa Hitachi Mita Hitachi City, Ibaraki Prefecture 5-2-1, Machi Ritsu Process Computer Engineering Co., Ltd. (72) Inventor Takayuki Yoneoka 5-2-1, Omika-cho, Hitachi-shi, Ibaraki Hirate Process Computer Engineering Co., Ltd. (72) Inventor Shinya Fujita 5-2-1 Omika-cho, Hitachi-shi, Ibaraki Prefecture Hitachi Process Computer Ta Engineering Co., Ltd. (72) Inventor Yasuo Kominato 5-2-1 Omika-cho, Hitachi City, Ibaraki Hitachi Process Computer Engineering Co., Ltd. (72) Inventor Toshiyuki Iwamoto 5-2-1 Omika-cho, Hitachi City, Ibaraki Prefecture No. Hitate Process Computer Engineering Co., Ltd. (72) Inventor Mikio Kuroki 5-2-1 Omika-cho, Hitachi City, Ibaraki Prefecture

Claims (4)

[Claims] 1. A transport device for moving an object to be measured having a barcode attached thereto, a plurality of image input devices installed on the transport device, and an image captured from the image input device. And a bar code recognition device including an image processing device that reads the bar code attached to the measurement object, filters images captured from the plurality of image input devices, and performs the filtering process. A barcode extraction unit that extracts the barcode from the image and the image captured from the image input device, a barcode scanning processing unit that extracts the feature amount of the extracted barcode, and a barcode scan processing unit based on the extracted feature amount. A barcode recognition device comprising a processing device having a barcode recognition unit for recognizing the barcode. 2. The barcode processing apparatus according to claim 1, wherein the processing apparatus includes a barcode inclination detection unit that detects an angle of the barcode extracted by the barcode extraction unit, and the barcode scan processing unit. A barcode recognition apparatus comprising: a processing unit that extracts a feature amount of the barcode based on the inclination of the barcode detected by a detection unit. 3. The bar according to claim 2, wherein the scan processing unit extracts a feature path of the barcode extracted by the barcode extraction unit and complements pixels of the barcode. Code recognition device. 4. A barcode which captures an image of an object to be measured to which a barcode is attached from an image device, processes the captured image, and reads the barcode attached to the object to be measured. In the recognition method, an image captured from the plurality of image input devices is filtered, the barcode is extracted from the filtered image and the image captured from the image input device, and the extracted barcode is Detecting the inclination and extracting the feature amount of the barcode according to the inclination,
A barcode recognition method for recognizing the barcode based on the extracted feature amount.