A tire barcode recognition method, system and device based on machine vision

Abstract

The present invention provides a tire barcode recognition method, system and device based on machine vision. The present invention is suitable for collecting tire images at a long distance on a tire conveyor line and quickly reading barcodes. The barcode is quickly located by first locating the tire and then locating the barcode, thereby improving the barcode recognition efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of machine vision recognition, and in particular to a tire barcode recognition method, system and device based on machine vision. Background Art

[0002] At present, the industry generally uses manual handheld scanners to scan and identify tire barcodes or uses barcode readers to read at close range. The former is labor-intensive, inefficient, and has high labor costs; the latter is not suitable for long-distance reading, and is easily affected by the distance of the barcode reader. In particular, it is difficult to accurately read the barcode under complex background light sources, resulting in a high scan miss rate, and is not suitable for large-scale industrial production. Summary of the invention

[0003] The problem solved by the present invention is how to replace manual work or a code reader to efficiently identify tire bar codes and reduce the probability of missed scanning.

[0004] To solve the above problems, the present invention provides a tire barcode recognition method based on machine vision, comprising the steps of:

[0005] Step 1: using an image collector to collect original tire images of tires on a tire conveyor line;

[0006] Step 2: preprocessing the original tire image;

[0007] Step 3: Detecting and locating the inner circle and outer circle of the tire from the preprocessed tire image;

[0008] Step 4: Locate and extract the barcode label between the inner circle and the outer circle of the tire;

[0009] Step 5: Locate and extract the barcode area from the barcode label;

[0010] Step 6: Read and decode the barcode in the barcode area;

[0011] Step 7: Verify the barcode decoding.

[0012] The beneficial effects of the present invention are as follows: the method is suitable for collecting tire images at a long distance on a tire conveyor line and quickly reading bar codes, wherein the bar codes are quickly located by first locating the tire and then locating the bar codes, thereby improving the bar code recognition efficiency.

[0013] Preferably, in step 1, an industrial high-definition area array camera is used to collect original tire images of the tire, and K original tire images are continuously taken each time, thereby improving the precision of the original tire images.

[0014] Preferably, step 2 comprises the following sub-steps:

[0015] Step 201, graying the tire image: graying the components of the R channel, the G channel, and the B channel at a preset ratio;

[0016] Gray=R*q+G*w+B*e

[0017] Among them, Gray is the gray value after transformation, q, w, and e are the preset ratios of R channel, G channel, and B channel respectively;

[0018] Step 202: Use a Gaussian filter to smooth the grayscale image.

[0019] Preferably, step 3 comprises the following sub-steps:

[0020] Step 301, setting M groups of thresholds of the Canny operator to extract M groups of tire image contour samples, each group of thresholds extracts K contour images, thereby obtaining M*K contour images;

[0021] Step 302: traverse the grayscale values ​​of all pixels on each contour map, and calculate the probability p that the grayscale values ​​of corresponding pixels in all contour maps are the same

[0022]

[0023] Where n is the number of contour images with the same grayscale value of pixels at the same position;

[0024] When p ≥ 85%, the pixel is retained; otherwise, the pixel is deleted and finally the contour image with complete edges is synthesized;

[0025] Step 303, presetting a first range value of the inner circle radius, using Hough transform to detect pixel points corresponding to the first range value in the contour map to obtain the inner circle edge of the tire, and locate the inner circle;

[0026] Step 304: using the center of the inner circle detected as the center of the outer circle, and presetting a second range value of the outer circle radius; using Hough transform to detect the outer circle within the second range value on the contour map; firstly, the Hough transform detects a partial arc of the outer circle, and the distance between the center and the arc is the radius of the outer circle, and then the Hough transform is used to fit and locate the outer circle according to the points on the arc, the center and the radius;

[0027] Before performing Hough transform, the Canny operator is used to synthesize the image to reduce the interference circles near the inner circle, thereby overcoming the problem of incomplete edge extraction caused by Hough transform, failure to detect circles or interference at the edges, and detection of multiple circles.

[0028] Preferably, after the inner circle edge is detected and located by using Hough transform in step 303, the pixel points of the inner circle located by Hough transform are subjected to circle property detection. If the pixel points meet the circle property, the pixel points are retained. Otherwise, the pixel points are excluded to locate the inner circle, which specifically includes:

[0029] ① In the inner circle edge detected by Hough transform, take t inner circle edge points with set intervals, and calculate the distance r between the t inner circle edge points and the center of the circle respectively. i (i=1, 2, 3, ..., t), and calculate r i The average Sequentially i and For comparison, if Then keep the inner circle edge point, if Then exclude the inner circle edge point and select the remaining s inner edge points that meet the properties of the circle;

[0030] ②Select two points from the s inner circle edge points, and the distance between the two inner circle edge points is Then take any remaining inner circle edge point B i (i=1,2,3,...,b) and the above two inner circle edge points form a triangle, and then determine whether the formed triangle is a right triangle. The determination formula is:

[0031] a 2 +b 2 =c 2

[0032] In the formula, a and b are two right-angled sides, and c is the diameter.

[0033] If a 2 +b 2 =c 2 , then a right triangle can be formed and the inner circle can be located; otherwise, determine whether the three sides of the triangle meet If satisfied, retain the inner circle edge point B i Otherwise, delete the inner circle edge point B i ; After screening, g inner edge points remain;

[0034] ③ Calculate the distance between any two inner edge points among the g inner edge points, and determine whether the distance between the two inner edge points is less than or equal to If the condition is satisfied, the corresponding inner edge point is retained; otherwise, the corresponding inner edge point is deleted; and finally, the inner edge point that meets the properties of the circle is screened out;

[0035] ④ Use Hough transform to fit the final filtered inner edge points and locate the inner circle.

[0036] By utilizing the properties of the circle to perform a secondary judgment on the located inner circle edge, the interference of the false inner circle edge is reduced.

[0037] Preferably, step 4 comprises the following sub-steps:

[0038] Step 401, binarize the original image, search and extract the area where the barcode label is located by using the block search method according to the inner circle and outer circle of the tire located in step 3, distinguish the tire tread and the barcode label by comparing the gray value, thereby determining the position of the barcode label, and extract the barcode label at the corresponding position in the original image, and improve the search efficiency of the barcode label by using the block search method;

[0039] Step 402: Perform image enhancement processing on the extracted barcode label.

[0040] Preferably, step 5 comprises the following sub-steps:

[0041] Step 501, barcode label preprocessing: use the median filter to remove noise from the barcode label extracted in step 4, and keep the edge of the barcode in the barcode label; then use the mean filter to eliminate the interference of the straight line in the barcode label that is not the edge of the barcode on the barcode positioning; then perform binarization processing on the barcode label;

[0042] Step 502, barcode area positioning: the barcode includes a barcode and a number, and the barcode area is positioned and extracted according to a set ratio range of the barcode and the number;

[0043] Step 503, barcode area position correction: the extracted barcode area is subjected to Radon transformation to obtain a projection image and data, and the inclination angle of the barcode area is obtained according to the projection image, and then the barcode area is rotated and corrected using a rotation algorithm;

[0044] Step 504: Use image dilation processing to repair defects at the edge of the barcode area, and erode the connected pixels between the barcodes in the barcode area;

[0045] Step 505: Determine whether the located barcode area is qualified according to the barcode area format standard.

[0046] Preferably, the step 7 comprises the sub-steps of:

[0047] Step 701, read the number decoded in step 6 from right to left, add the numbers of the even digits from the second digit and multiply by 3 to get the previous value; then add the numbers of the odd digits from the third digit to get the next value, and then add the previous value and the next value to get N;

[0048] Step 702: Subtract the unit digit of N from 10 to obtain C, which is the check value;

[0049] C = 10-N mod 10;

[0050] Step 703, determine whether C is equal to the first digit in step 701; if they are equal, the barcode is read correctly; otherwise, the barcode is read incorrectly, and return to step 6.

[0051] A tire barcode recognition system based on machine vision includes an image acquisition module for acquiring an original tire image, an image preprocessing module for graying and smoothing the original tire image, a first positioning module for detecting and positioning the inner circle and outer circle of the tire, a second positioning module for positioning the barcode label, a barcode extraction module for locating and extracting the barcode area from the barcode label, a reading module for reading and decoding the extracted barcode, and a verification module for verifying the decoded barcode, wherein:

[0052] The image preprocessing module comprises a grayscale processing unit and a smoothing processing unit connected in sequence, and the grayscale processing unit is connected to the image acquisition module;

[0053] The first positioning module comprises an inner circle positioning unit and an outer circle positioning unit connected in sequence, and the inner circle positioning unit is connected to the smoothing processing unit;

[0054] The second positioning module is also connected to the image acquisition module;

[0055] The barcode extraction module includes a barcode label preprocessing unit, a barcode area positioning unit, a barcode area position correction unit, a barcode area repair unit and a barcode area verification unit which are connected in sequence, and the barcode label preprocessing unit is connected to the second positioning module;

[0056] The reading module is connected to the barcode area verification unit.

[0057] A computing device comprising:

[0058] One or more processors, a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include methods for executing the tire barcode recognition method based on machine vision as described above. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 is a flow chart;

[0060] Figure 2 is the original tire image;

[0061] Figure 3 This is the effect picture after the image is grayed out;

[0062] Figure 4 This is the effect picture after Gaussian filtering smoothing;

[0063] Figure 5 It is the overall outline of the synthesis;

[0064] Figure 6 This is the effect of binarization of the original tire image.

[0065] Figure 7 Schematic diagram of block search for barcode labels. DETAILED DESCRIPTION

[0066] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0067] like Figure 1 As shown, a tire barcode recognition method based on machine vision includes the following steps:

[0068] Step 1, using an image collector to collect original tire images of tires on the tire conveyor line; this specific embodiment uses a light shielding box with a height of 1m set on the tire conveyor line, the light shielding box is an identification area, and a photoelectric sensor is set at the identification area. The image collector is an industrial high-definition area array camera, which is set at the identification area to capture the actual tire and barcode on the tire conveyor line. When the photoelectric sensor senses that the tire is conveyed to the identification area, it triggers the high-definition area array camera to collect the original tire image, and the high-definition area array camera continuously takes K original tire images each time;

[0069] like Figure 2 As shown in the figure, in the original tire image, the background of the inner circle of the tire is complex, containing a large number of circles, and the color of the inner circle outline is significantly different from the metal sheet in the background. The outer circle is also different in color from the background, but there are black quadrilaterals at the four corners, and their colors are similar to the tire color. After zooming in on the image, you can see the barcode. Due to the small area of ​​the barcode itself and the long shooting distance, the barcode image is blurred and the angle is not correct;

[0070] Step 2: preprocessing the original tire image;

[0071] Step 201, graying the tire image: graying the components of the R channel, the G channel, and the B channel at a preset ratio;

[0072] Gray=R*q+G*w+B*e

[0073] Among them, Gray is the gray value after transformation, q, w, and e are the preset ratios of R channel, G channel, and B channel respectively, such as Figure 3 As shown;

[0074] Step 202: Use a Gaussian filter to smooth the grayscale image. Figure 4 shown.

[0075] Step 3: Detecting and locating the inner circle and outer circle of the tire from the preprocessed tire image;

[0076] Step 301, setting M groups of thresholds of the Canny operator to extract M groups of tire image contour samples, each group of thresholds extracts K contour images, thereby obtaining M*K contour images;

[0077] Step 302: traverse the grayscale values ​​of all pixels on each contour map, and calculate the probability p that the grayscale values ​​of corresponding pixels in all contour maps are the same

[0078]

[0079] Where n is the number of contour images with the same grayscale value of pixels at the same position;

[0080] When p ≥ 85%, the pixel is retained; otherwise, the pixel is deleted, and finally the contour image with complete edges is synthesized, such as Figure 5 As shown; the inner circle is completed by the Canny operator to reduce the interference around the inner circle;

[0081] Step 303, presetting a first range value of the inner circle radius, using Hough transform to detect pixel points corresponding to the first range value in the contour map to obtain the inner circle edge of the tire, and locate the inner circle;

[0082] Step 304: using the center of the inner circle detected as the center of the outer circle, and presetting a second range value of the outer circle radius; using Hough transform to detect the outer circle within the second range value on the contour map; firstly, the Hough transform detects a partial arc of the outer circle, and the distance between the center and the arc is the radius of the outer circle, and then the Hough transform is used to fit and locate the outer circle according to the points on the arc, the center and the radius;

[0083] Before performing Hough transform, the Canny operator is used to synthesize the image to reduce the interference circles near the inner circle, thereby overcoming the problem of incomplete edge extraction caused by Hough transform, inability to detect circles or interference at the edges, and detecting multiple circles;

[0084] Step 4: Locate and extract the barcode label between the inner circle and the outer circle of the tire;

[0085] Step 401: Figure 6 As shown, the original image is binarized, and the area where the barcode label is located is obtained according to the inner circle and outer circle of the tire located in step 3. The tire tread and the barcode label are distinguished by comparing the grayscale values, as shown in FIG. Figure 7 As shown, the block search method is used to Figure 6Divide into 4 parts, search from left to right and from top to bottom to determine the position of the barcode label, and extract the barcode label at the corresponding position in the original image, and improve the search efficiency of the barcode label through the block search method;

[0086] Step 402: Perform image enhancement processing on the extracted barcode label.

[0087] Step 5: Locate and extract the barcode area from the barcode label;

[0088] Step 501, barcode label preprocessing: use the median filter to remove noise from the barcode label extracted in step 4, and keep the edge of the barcode in the barcode label; then use the mean filter to eliminate the interference of the straight line in the barcode label that is not the edge of the barcode on the barcode positioning; then perform binarization processing on the barcode label;

[0089] Step 502, barcode area positioning: the barcode includes a barcode and a number, and the barcode area is positioned and extracted according to a set ratio range of the barcode and the number;

[0090] Step 503, barcode area position correction: the extracted barcode area is subjected to Radon transformation to obtain a projection image and data, and the inclination angle of the barcode area is obtained according to the projection image, and then the barcode area is rotated and corrected using a rotation algorithm;

[0091] Step 504: Use image dilation processing to repair defects at the edge of the barcode area, and erode the connected pixels between the barcodes in the barcode area;

[0092] Step 505: judging whether the located barcode area is qualified according to the barcode area format standard;

[0093] Among them, the format standard of the barcode area is:

[0094] Whether the width and height ratio of the barcode area meets the preset standards;

[0095] Whether the number of barcodes in the barcode area meets the preset number of barcodes.

[0096] Step 6: Read and decode the barcode in the barcode area; decoding the barcode is a prior art and will not be described in detail here;

[0097] Step 7: Verify the decoding of the barcode;

[0098] Step 701, read the number decoded in step 6 from right to left, add the numbers of the even digits from the second digit and multiply by 3 to get the previous value; then add the numbers of the odd digits from the third digit to get the next value, and then add the previous value and the next value to get N;

[0099] Step 702: Subtract the unit digit of N from 10 to obtain C, which is the check value;

[0100] C = 10-N mod 10;

[0101] Step 703, determine whether C is equal to the first digit in step 701; if they are equal, the barcode is read correctly; otherwise, the barcode is read incorrectly, and return to step 6.

[0102] In addition, after the inner circle edge is detected and located using Hough transform in step 303, the pixel points of the inner circle located by Hough transform are subjected to circle property detection. If the pixel points meet the circle property, the pixel points are retained. Otherwise, the pixel points are excluded to locate the inner circle. Specifically, the following steps are performed:

[0103] ① In the inner circle edge detected by Hough transform, take t inner circle edge points with set intervals, and calculate the distance r between the t inner circle edge points and the center of the circle respectively. i (i=1, 2, 3, ..., t), and calculate r i The average Sequentially i and For comparison, if Then keep the inner circle edge point, if Then exclude the inner circle edge point and select the remaining s inner edge points that meet the properties of the circle;

[0104] ② Select two points from the s inner circle edge points, and the distance between the two inner circle edge points is 2r; then take any remaining inner circle edge point B i (i=1,2,3,...,b) and the above two inner circle edge points form a triangle, and then determine whether the formed triangle is a right triangle. The determination formula is:

[0105] a 2 +b 2 =c 2

[0106] In the formula, a and b are two right-angled sides, and c is the diameter.

[0107] If a 2 +b 2 =c 2 , then a right triangle can be formed and the inner circle can be located; otherwise, determine whether the three sides of the triangle meet If satisfied, retain the inner circle edge point B i Otherwise, delete the inner circle edge point B i ; After screening, g inner edge points remain;

[0108] ③ Calculate the distance between any two inner edge points among the g inner edge points, and determine whether the distance between the two inner edge points is less than or equal to If the condition is satisfied, the corresponding inner edge point is retained; otherwise, the corresponding inner edge point is deleted; and finally, the inner edge point that meets the properties of the circle is screened out;

[0109] ④ Use Hough transform to fit the final filtered inner edge points and locate the inner circle.

[0110] By utilizing the properties of the circle to perform a secondary judgment on the located inner circle edge, the interference of the false inner circle edge is reduced. Specific embodiment 2

[0112] A tire barcode recognition system based on machine vision includes an image acquisition module for acquiring an original tire image, an image preprocessing module for graying and smoothing the original tire image, a first positioning module for detecting and positioning the inner circle and outer circle of the tire, a second positioning module for positioning the barcode label, a barcode extraction module for locating and extracting the barcode area from the barcode label, a reading module for reading and decoding the extracted barcode, and a verification module for verifying the decoded barcode, wherein:

[0113] The image preprocessing module comprises a grayscale processing unit and a smoothing processing unit connected in sequence, and the grayscale processing unit is connected to the image acquisition module;

[0114] The first positioning module comprises an inner circle positioning unit and an outer circle positioning unit connected in sequence, and the inner circle positioning unit is connected to the smoothing processing unit;

[0115] The second positioning module is also connected to the image acquisition module;

[0116] The barcode extraction module includes a barcode label preprocessing unit, a barcode area positioning unit, a barcode area position correction unit, a barcode area repair unit and a barcode area verification unit which are connected in sequence, and the barcode label preprocessing unit is connected to the second positioning module;

[0117] The reading module is connected to the barcode area verification unit. Specific embodiment 3

[0119] A computing device comprising:

[0120] One or more processors, a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include a method for executing the machine vision-based tire barcode recognition method described in Specific Embodiment 1.

Claims

1. A tire barcode recognition method based on machine vision, characterized in that: Includes steps: Step 1: using an image collector to collect original tire images of tires on a tire conveyor line; Step 2: preprocessing the original tire image; Step 3: Detecting and locating the inner circle and outer circle of the tire from the preprocessed tire image; The method comprises the following sub-steps: step 301, setting M groups of thresholds of the Canny operator to extract contour samples of M groups of tire images, extracting K contour images for each group of thresholds, thereby obtaining M*K contour images; Step 302: traverse the grayscale values ​​of all pixels on each contour map, and calculate the probability p that the grayscale values ​​of corresponding pixels in all contour maps are the same Where n is the number of contour images with the same grayscale value of pixels at the same position; When p ≥ 85%, the pixel is retained; Otherwise, delete the pixel points and finally synthesize the overall contour image with complete edges; Step 303, presetting a first range value of the inner circle radius, using Hough transform to detect pixel points corresponding to the first range value in the contour map to obtain the inner circle edge of the tire, and locate the inner circle; Step 304: using the center of the inner circle detected as the center of the outer circle, and presetting a second range value of the outer circle radius; using Hough transform to detect the outer circle within the second range value on the contour map; firstly, the Hough transform detects a partial arc of the outer circle, and the distance between the center and the arc is the radius of the outer circle, and then the Hough transform is used to fit and locate the outer circle according to the points on the arc, the center and the radius; Step 4: Locate and extract the barcode label between the inner circle and the outer circle of the tire; Includes sub-steps: Step 401, binarizing the original image, searching and extracting the area where the barcode label is located by using a block search method based on the inner circle and outer circle of the tire located in step 3, distinguishing the tire tread and the barcode label by comparing the grayscale values, thereby determining the position of the barcode label, and extracting the barcode label at the corresponding position in the original image; Step 402: Perform image enhancement processing on the extracted barcode label; Step 5: Locate and extract the barcode area from the barcode label; Step 6: Read and decode the barcode in the barcode area; Step 7: Verify the barcode decoding.

2. The tire barcode recognition method based on machine vision according to claim 1, characterized in that: In the step 1, an industrial high-definition area array camera is used to collect original tire images of the tire, and K original tire images are continuously taken each time.

3. The tire barcode recognition method based on machine vision according to claim 1, characterized in that: The step 2 includes the following sub-steps: Step 201, graying the tire image: graying the components of the R channel, the G channel, and the B channel at a preset ratio; Gray=R*q+G*w+B*e Among them, Gray is the gray value after transformation, q, w, and e are the preset ratios of R channel, G channel, and B channel respectively; Step 202: Use a Gaussian filter to smooth the grayscale image.

4. The tire barcode recognition method based on machine vision according to claim 1, characterized in that: After the inner circle edge is detected and located using Hough transform in step 303, the pixel points of the inner circle located by Hough transform are subjected to circle property detection. If the pixel points meet the circle property, the pixel points are retained. Otherwise, the pixel points are excluded to locate the inner circle. Specifically, the following steps are performed: ① In the inner circle edge detected by Hough transform, take t inner circle edge points with set intervals, and calculate the distance r between the t inner circle edge points and the center of the circle respectively. i (i = 1, 2, 3, ..., t), and calculate r i The average Sequentially i and For comparison, if Then keep the inner circle edge point, if Then exclude the inner circle edge point and select the remaining s inner edge points that meet the properties of the circle; ②Select two points from the s inner circle edge points, and the distance between the two inner circle edge points is Then take any remaining inner circle edge point B i (i=1,2,3,...,b) and the above two inner circle edge points form a triangle, and then determine whether the formed triangle is a right triangle. The determination formula is: a 2 +b 2 =c 2 In the formula, a and b are two right-angled sides, and c is the diameter. If a 2 +b 2 =c 2 , then a right triangle can be formed and the inner circle can be located; otherwise, determine whether the three sides of the triangle meet If satisfied, retain the inner circle edge point B i Otherwise, delete the inner circle edge point B i ; After screening, g inner edge points remain; ③ Calculate the distance between any two inner edge points among the g inner edge points, and determine whether the distance between the two inner edge points is less than or equal to If the condition is satisfied, the corresponding inner edge point is retained; otherwise, the corresponding inner edge point is deleted; and finally, the inner edge point that meets the properties of the circle is screened out; ④ Use Hough transform to fit the final filtered inner edge points and locate the inner circle.

5. The tire barcode recognition method based on machine vision according to claim 1, characterized in that: The step 5 comprises the following sub-steps: Step 501, barcode label preprocessing: use median filtering to denoise the barcode label extracted in step 4, and keep the edge of the barcode in the barcode label; then use mean filtering to eliminate the interference of the straight line in the barcode label that is not the edge of the barcode on the barcode positioning; then binarize the barcode label; Step 502, barcode area positioning: the barcode includes the barcode and the number, and the barcode area is positioned and extracted according to the set ratio range of the barcode and the number; Step 503, barcode area position correction: the extracted barcode area is subjected to Radon transformation to obtain a projection image and data, and the inclination angle of the barcode area is obtained according to the projection image, and then the barcode area is rotated and corrected using a rotation algorithm; Step 504: Use image dilation processing to repair defects at the edge of the barcode area, and erode the connected pixels between the barcodes in the barcode area; Step 505: Determine whether the located barcode area is qualified according to the barcode area format standard.

6. The tire barcode recognition method based on machine vision according to claim 1, characterized in that: The step 7 comprises the following sub-steps: Step 701, read the number decoded in step 6 from right to left, add the numbers of the even digits from the second digit and multiply by 3 to get the previous value; then add the numbers of the odd digits from the third digit to get the next value, and then add the previous value and the next value to get N; Step 702: Subtract the unit digit of N from 10 to obtain C, which is the check value; C = 10-N mod 10; Step 703, determine whether C is equal to the first digit in step 701; if they are equal, the barcode is read correctly; otherwise, the barcode is read incorrectly, and return to step 6.

7. A tire barcode recognition system based on machine vision, based on the tire barcode recognition method based on machine vision according to any one of claims 1 to 6, characterized in that: The system comprises an image acquisition module for acquiring an original tire image, an image preprocessing module for graying and smoothing the original tire image, a first positioning module for detecting and positioning the inner circle and outer circle of the tire, a second positioning module for positioning the barcode label, a barcode extraction module for locating and extracting the barcode area from the barcode label, a reading module for reading and decoding the extracted barcode, and a verification module for verifying the decoded barcode, wherein: The image preprocessing module comprises a grayscale processing unit and a smoothing processing unit connected in sequence, and the grayscale processing unit is connected to the image acquisition module; The first positioning module comprises an inner circle positioning unit and an outer circle positioning unit connected in sequence, and the inner circle positioning unit is connected to the smoothing processing unit; The second positioning module is also connected to the image acquisition module; The barcode extraction module includes a barcode label preprocessing unit, a barcode area positioning unit, a barcode area position correction unit, a barcode area repair unit and a barcode area verification unit which are connected in sequence, and the barcode label preprocessing unit is connected to the second positioning module; The reading module is connected to the barcode area verification unit.

8. A computing device, characterized in that include: One or more processors, a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include a method for executing the tire barcode recognition method based on machine vision as described in any one of claims 1 to 6.

Images