A light guide plate hot-pressing line post-AOI detection system and detection method

Abstract

The application discloses a kind of light guide plate hot-pressing line after AOI detection system and detection method, it is related to light guide plate detection technical field, comprising: image acquisition module, gray processing module, binarization module, outline extraction module, reference point setting module, detection value acquisition module, threshold acquisition module and judging module;The reference point setting module is used to obtain detection dot reference point based on dot outline;The detection value acquisition module is used to obtain dot detection value based on detection dot outline and detection dot reference point;The threshold acquisition module is used to obtain first qualified threshold and second qualified threshold based on the image of first quantity normal light guide plate after hot-pressing;The judging module is used to judge whether light guide plate is qualified after hot-pressing based on dot detection value, first qualified threshold and second qualified threshold, the present application is used to solve the problem that existing light guide plate detection technology cannot analyze dot structure, leading to inaccurate qualified detection after light guide plate hot-pressing.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of light guide plate detection, in particular to a light guide plate heat pressing line after AOI detection system and method. BACKGROUND

[0002] The main function of the light guide plate is to uniformly distribute the light emitted by the light source to the entire display panel, thereby realizing uniform brightness display. In the manufacturing process of the light guide plate, the heat pressing line process is a common technical means for forming dot structures on the surface of the light guide plate. These dot structures can effectively scatter light and improve the optical performance of the light guide plate. The heat pressing line process needs to accurately control the temperature and pressure. If the temperature and pressure deviate, the obtained dot structure will also deviate, resulting in unqualified light guide plates.

[0003] The existing light guide plate detection directly detects whether the light guide plate contains scratches, dirt and damage and the like. However, after the heat pressing of the light guide plate, it is also necessary to judge whether the dot structure of the light guide plate is qualified. The detection of the dot structure is different from the detection of scratches, dirt and damage and the like. If the dot structure is unqualified, it will cause the scattered light of the light guide plate to decrease, affecting the normal use of the light guide plate. For example, in the patent application with the publication number CN114881987A, a heat-pressed light guide plate defect visualization detection method based on improved YOLOv5 is disclosed. This scheme directly detects whether the light guide plate contains scratches, dirt and damage and the like, and fails to detect the dot structure, resulting in the light guide plate still containing defects. That is, the existing light guide plate detection technology fails to analyze the dot structure, resulting in inaccurate qualified detection after the heat pressing of the light guide plate. SUMMARY

[0004] The present application aims to at least solve one of the technical problems in the prior art. By obtaining a detection dot reference point based on a dot contour, obtaining a dot detection value based on the detection dot contour and the detection dot reference point, obtaining a first qualified threshold and a second qualified threshold based on images of a first number of normal light guide plates after heat pressing, and judging whether the light guide plate is qualified after heat pressing based on the dot detection value, the first qualified threshold and the second qualified threshold, the problem of inaccurate qualified detection after the heat pressing of the light guide plate due to the failure of the existing light guide plate detection technology to analyze the dot structure is solved.

[0005] To achieve the above-mentioned purpose, the present application provides a light guide plate heat pressing line after AOI detection system, which comprises:

[0006] An image acquisition module, a gray scale processing module, a binaryzation module, a contour extraction module, a reference point setting module, a detection value acquisition module, a threshold value acquisition module and a judgment module;

[0007] The image acquisition module is used to acquire an image of the light guide plate after heat pressing, which is marked as a heat-pressed image.

[0008] The grayscale processing module is used to perform grayscale processing on the hot-pressed image to obtain a grayscale image after hot pressing.

[0009] The binarization module is used to binarize the hot-pressed grayscale image to obtain the binarized image after hot pressing.

[0010] The contour extraction module is used to obtain the contour of the detected dots based on the binarized image after hot pressing.

[0011] The reference point setting module is used to obtain reference points for detected dots based on the dot contour.

[0012] The detection value acquisition module is used to acquire the detection value of the detection point based on the detection point outline and the detection point reference point;

[0013] The threshold acquisition module is used to acquire a first qualified threshold and a second qualified threshold based on the image of the first normal light guide plate after hot pressing.

[0014] The judgment module is used to determine whether the light guide plate is qualified after hot pressing based on the dot detection value, the first qualified threshold and the second qualified threshold.

[0015] Furthermore, the grayscale processing module is configured with a grayscale processing strategy, which includes:

[0016] Obtain the R, G, and B values ​​of any pixel in the hot-pressed image and label them as the detected R value, detected G value, and detected B value, respectively. Calculate the mean of the detected R, G, and B values ​​and label it as the detected grayscale value. Replace the detected R, G, and B values ​​with the detected grayscale value, and represent each pixel in the hot-pressed image with the corresponding detected grayscale value to obtain the hot-pressed grayscale image.

[0017] Furthermore, the binarization configuration includes a binarization strategy, which comprises:

[0018] Divide the grayscale values ​​from 0 to 255 into a equal parts, labeled as equal parts; the length of each equal part is 255÷a; the starting value and ending value of each equal part are (255÷a)×b and (255÷a)×(b+1), respectively; where b is a positive integer from 0 to a-1.

[0019] Obtain the number of grayscale values ​​detected in each evenly divided region and mark it as the number of detections;

[0020] A histogram is plotted with grayscale values ​​on the X-axis, the number of detections on the Y-axis, and the evenly divided portions as histogram intervals. This histogram is then labeled as the detection grayscale histogram.

[0021] The portion of the grayscale histogram that is greater than the number of adjacent detections on the left and right sides is marked as the first target portion;

[0022] The first target portion with the smallest initial value is identified and marked as the second target portion.

[0023] Obtain the first target portion with the largest initial value and mark it as the third target portion;

[0024] The part with the smallest initial value between the second and third target parts is identified and marked as the fourth target part.

[0025] Calculate the mean of the starting and ending values ​​of the fourth target part and mark them as the binarization threshold;

[0026] In the grayscale image after hot pressing, the detected grayscale values ​​greater than the halftone dot binarization threshold are set to 0, and the detected grayscale values ​​less than or equal to the binarization threshold are set to 255 to obtain the binarized image after hot pressing.

[0027] Furthermore, the contour extraction is configured with a contour extraction strategy, which includes:

[0028] In the binarized image after hot pressing, obtain the pixel with a detected gray value of 0 that is adjacent to the detected gray value of 255 and mark it as the initial pixel; determine whether there are pixels with a detected gray value of 0 around the initial pixel. If not, obtain a new initial pixel; if so, mark the initial pixel as the edge starting pixel.

[0029] Starting with the pixel adjacent to the left of the edge starting pixel, and using the edge starting pixel as the rotation center, rotate clockwise to search for the first pixel whose grayscale value changes from 255 to 0, and mark it as the edge search pixel; then use the edge search pixel as the rotation center to obtain a new edge search pixel, and stop the operation when the new edge search pixel is the same as the edge starting pixel, and obtain the contour composed of the edge starting pixel and all edge search pixels, and mark it as the detection dot contour.

[0030] Furthermore, the reference point setting configuration includes a reference point setting strategy, which comprises:

[0031] Obtain the outlines of all detection dots in the binarized image after hot pressing;

[0032] Draw a second number of coordinate points at equal intervals on the outline of the detected dots, and mark them as outline coordinate points;

[0033] Connect the contour coordinate points sequentially along the contour of the detection points to obtain line segments, which are marked as contour connection line segments;

[0034] Obtain the included angle between all adjacent contour connecting line segments and mark it as the contour angle;

[0035] Calculate the mean of the contour angles and mark it as the mean of the contour angles;

[0036] Calculate the absolute value of the difference between all contour angles and the mean of the contour angles, and mark it as the reference difference;

[0037] Sort the reference differences from smallest to largest, obtain the intersection points of the contour angles corresponding to the first four reference differences, and mark them as the first coordinate point, the second coordinate point, the third coordinate point, and the fourth coordinate point, respectively.

[0038] Furthermore, the reference point setting module is configured with a reference point setting strategy, which includes:

[0039] Connect the first coordinate point and the second coordinate point to obtain a line segment, which is marked as the first line segment; connect the second coordinate point and the third coordinate point to obtain a line segment, which is marked as the second line segment;

[0040] Draw a perpendicular line from the midpoint of the first line segment to the first line segment, and mark it as the first perpendicular line; draw a perpendicular line from the midpoint of the second line segment to the second line segment, and mark it as the second perpendicular line.

[0041] Obtain the intersection of the first and second perpendicular lines and mark it as the reference point for the detection network.

[0042] Furthermore, the detection value acquisition module is configured with a detection value acquisition strategy, which includes:

[0043] Mark the pixels on the detected dot outline as detected outline pixels;

[0044] Obtain the distance from the reference point of the detected halftone dot to the pixel point of the detected contour, and mark it as the halftone dot detection value.

[0045] Furthermore, the threshold acquisition module is configured with a threshold acquisition strategy, which includes:

[0046] The image after the first normal number of light guide plates are hot-pressed is marked as a normal image; the detection values ​​of all the dots in the outline of each detection dot in the normal image are obtained and marked as normal distances;

[0047] Obtain the range of normal distances, labeled C1 to Cn; divide the range of fall comparison values ​​evenly into d parts, labeled as normal parts; the length of each normal part is (Cn-C1)÷d; the starting and ending values ​​of each normal part are C1+[(Cn-C1)÷d]×e and C1+[(Cn-C1)÷d]×(e+1), respectively; where e is a positive integer from 0 to d-1.

[0048] Count the number of each normal part and mark it as the number of normal parts;

[0049] Sort the number of normal parts in ascending order of their starting values ​​to obtain normal sequence numbers, which are positive integers starting from 1;

[0050] Obtain the frequency of normal distances and label it Z;

[0051] The threshold for the number of normal cases is calculated as: G = f × (Z ÷ d); where Z is the threshold for classifying falls, and f is the ratio from 0 to 1.

[0052] The system checks whether the number of normal parts is less than the normal number threshold according to the normal sequence number from smallest to largest. If not, the check stops. If so, the number of normal parts is deleted, and the system continues to check whether the number of normal parts is less than the normal number threshold until the number of normal parts is greater than or equal to the normal number threshold. The system then obtains the starting value of the normal part corresponding to the number of normal parts on the leftmost side and marks it as the first qualified threshold.

[0053] The system checks whether the number of normal parts is less than the normal number threshold in descending order of normal sequence number. If not, the check stops. If so, the number of normal parts is deleted, and the system continues to check whether the number of normal parts is less than the normal number threshold until the number of normal parts is greater than or equal to the normal number threshold. The system then obtains the end value of the normal part corresponding to the number of normal parts on the far right and marks it as the second qualified threshold.

[0054] Furthermore, the judgment module is configured with a judgment strategy, which includes:

[0055] Determine whether the detection values ​​of all points are greater than or equal to the first qualified threshold and less than or equal to the second qualified threshold. If yes, generate a signal that the heat pressing of the light guide plate is qualified; otherwise, generate a signal that the heat pressing of the light guide plate is unqualified.

[0056] This invention provides an AOI inspection method for a light guide plate after hot pressing, comprising the following steps: acquiring an image of the light guide plate after hot pressing and marking it as a hot-pressed image;

[0057] The image after hot pressing is processed into grayscale to obtain a grayscale image after hot pressing.

[0058] Binarize the hot-pressed grayscale image to obtain the binarized image after hot pressing;

[0059] The outline of the detection dots is obtained based on the binarized image after hot pressing;

[0060] Reference points for detected dots are obtained based on dot contours;

[0061] The detection values ​​of the detection points are obtained based on the outline of the detection points and the reference points of the detection points.

[0062] The first and second qualified thresholds are obtained based on the images of the first number of normal light guide plates after hot pressing.

[0063] The quality of the light guide plate after hot pressing is determined based on the grid detection value, the first qualified threshold, and the second qualified threshold.

[0064] The beneficial effects of this invention are as follows: This invention obtains reference points for detected dots based on dot contours; obtains dot detection values ​​based on the dot contours and reference points; obtains a first and a second pass threshold based on the image of a light guide plate after hot pressing with a first normal quantity of dots; and determines whether the light guide plate is qualified after hot pressing based on the dot detection values, the first and second pass thresholds. The advantage is that it analyzes the dot structure, thereby improving the accuracy of the light guide plate's qualification detection.

[0065] This invention obtains reference points for the detection dots based on the dot contour; and obtains the dot detection values ​​based on the detection dot contour and the detection dot reference points. The advantage is that it can digitize the structure of the dots, making it convenient to judge whether the light guide plate is qualified based on the dots, while also improving the accuracy of the light guide plate's qualification test. Attached Figure Description

[0066] Figure 1 This is a schematic diagram of the system of the present invention;

[0067] Figure 2 This is a schematic diagram illustrating the detection of grayscale histograms according to the present invention;

[0068] Figure 3 This is a schematic diagram of the included angle of the contour of the present invention;

[0069] Figure 4 This is a schematic diagram of the reference points for the detection network of the present invention;

[0070] Figure 5 This is a flowchart of the steps of the present invention. Detailed Implementation

[0071] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0072] Example 1, please refer to Figure 1As shown, an AOI inspection system for a light guide plate after hot pressing includes: an image acquisition module, a grayscale processing module, a binarization module, a contour extraction module, a reference point setting module, a detection value acquisition module, a threshold acquisition module, and a judgment module.

[0073] The image acquisition module is used to acquire images of the light guide plate after hot pressing and mark them as hot-pressed images. During image acquisition, light is shone from the side to make the dots appear, and the background is set to black to make the dots easier to distinguish.

[0074] The grayscale processing module is used to perform grayscale processing on the hot-pressed image to obtain a grayscale image after hot pressing;

[0075] The grayscale processing module is configured with grayscale processing strategies, which include:

[0076] Obtain the R, G, and B values ​​of any pixel in the hot-pressed image and label them as the detected R value, detected G value, and detected B value, respectively. Calculate the mean of the detected R, G, and B values ​​and label it as the detected grayscale value. Replace the detected R, G, and B values ​​with the detected grayscale value, and represent each pixel in the hot-pressed image with the corresponding detected grayscale value to obtain the hot-pressed grayscale image.

[0077] In practical applications, the R, G, and B values ​​of any pixel in the hot-pressed image are obtained, which are 206, 212, and 224 respectively. The mean of 206, 212, and 224 is calculated to be 214, so the detected grayscale value is 214. The calculation result is kept as an integer.

[0078] The binarization module is used to binarize the hot-pressed grayscale image to obtain the binarized image after hot pressing.

[0079] Binarization configuration includes binarization strategies, which include:

[0080] The grayscale values ​​from 0 to 255 are evenly divided into 'a' parts, which are marked as equal parts; the length of each equal part is 255 ÷ a; the starting value and ending value of each equal part are (255 ÷ a) × b and (255 ÷ a) × (b + 1) respectively; where b is a positive integer from 0 to a-1; the even parts are divided to observe the distribution of detected grayscale values, for example, a is set to 8;

[0081] In practical applications, the gray values ​​from 0 to 255 are evenly divided into 8 parts; the length of the 8 equal parts is 255÷8; when b is 0, the starting value and ending value of each equal part are (255÷8)×0=0 and (255÷8)×(0+1)=31 respectively. The calculation results are rounded to the nearest integer, so an equal part is from 0 to 31.

[0082] Obtain the number of grayscale values ​​detected in each evenly divided region and mark it as the number of detections;

[0083] A histogram is plotted with grayscale values ​​on the X-axis, the number of detections on the Y-axis, and the evenly divided portions as histogram intervals. This histogram is then labeled as the detection grayscale histogram.

[0084] For practical applications, please refer to Figure 2 As shown, the detection grayscale histogram was plotted.

[0085] The portion of the grayscale histogram that is greater than the number of adjacent detections on the left and right sides is marked as the first target portion;

[0086] The first target portion with the smallest initial value is identified and marked as the second target portion.

[0087] Obtain the first target portion with the largest initial value and mark it as the third target portion;

[0088] The part with the smallest initial value between the second and third target parts is identified and marked as the fourth target part.

[0089] Calculate the mean of the starting and ending values ​​of the fourth target part and mark them as the binarization threshold;

[0090] In the grayscale image after hot pressing, the detected grayscale values ​​greater than the halftone dot binarization threshold are set to 0, and the detected grayscale values ​​less than or equal to the binarization threshold are set to 255 to obtain the binarized image after hot pressing. Because the background is black and the halftone dots have high light refraction, some grayscale values ​​of the halftone dots are higher than the background, so binarization can be performed.

[0091] For practical applications, please refer to Figure 2 As shown, the second target portion is 63 to 95, the third target portion is 191 to 223, and the average portion with the smallest starting value between the second and third target portions is 127 to 159. Therefore, the average of the starting and ending values ​​of the fourth target portion is (127 + 159) ÷ 2 = 143. Thus, the binarization threshold is 143. In the grayscale image after hot pressing, the detection grayscale values ​​greater than 143 are set to 0, and the detection grayscale values ​​less than or equal to 143 are set to 255 to obtain the binarized image after hot pressing.

[0092] The contour extraction module is used to obtain the contour of the detected dots based on the binarized image after hot pressing;

[0093] Contour extraction is configured with contour extraction strategies, which include:

[0094] In the binarized image after hot pressing, the pixel with a detected gray value of 0 adjacent to the detected gray value of 255 is obtained and marked as the initial pixel; it is determined whether there are pixels with a detected gray value of 0 around the initial pixel. If not, a new initial pixel is obtained; if so, the initial pixel is marked as the edge starting pixel. Since the light guide point is not an isolated pixel, this method can exclude isolated pixels and improve the recognition of the light guide point.

[0095] Starting with the pixel adjacent to the left of the edge starting pixel, and using the edge starting pixel as the rotation center, rotate clockwise to search for the first pixel whose grayscale value changes from 255 to 0, and mark it as the edge search pixel. Then, using the edge search pixel as the rotation center, obtain a new edge search pixel, and stop the operation when the new edge search pixel is the same as the edge starting pixel. Obtain the contour composed of the edge starting pixel and all edge search pixels, and mark it as the detection dot contour. Use the contour search method to obtain the contour of each light guide point.

[0096] The reference point setting module is used to obtain reference points for detected dots based on the dot outline;

[0097] The reference point setting configuration includes a reference point setting strategy, which includes:

[0098] Obtain the outlines of all detection dots in the binarized image after hot pressing;

[0099] Draw a second number of coordinate points at equal intervals on the outline of the detected dots, and mark them as outline coordinate points;

[0100] Connect the contour coordinate points sequentially along the contour of the detection points to obtain line segments, which are marked as contour connection line segments;

[0101] Obtain the included angle between all adjacent contour connecting line segments and mark it as the contour angle;

[0102] Calculate the mean of the contour angles and mark it as the mean of the contour angles;

[0103] Calculate the absolute value of the difference between all contour angles and the mean of the contour angles, and mark it as the reference difference;

[0104] The beneficial effect of obtaining the reference difference is as follows: Since the dots of the light guide plate are circular, the reference difference should be equal. However, the actual production dot is not a perfectly regular circle. At the same time, when there are defects in the dots of the light guide plate, there will be incomplete circles. By using a larger reference difference, it can be determined whether the contour angle is too large or too small. If the contour angle is too large or too small, it may be due to the incomplete part of the dots of the light guide plate. Here, in order to obtain dots that exclude the incomplete parts, the reference points of the dots obtained in the subsequent inspection will be more accurate.

[0105] Sort the reference differences from smallest to largest, obtain the intersection points of the contour angles corresponding to the first four reference differences, and mark them as the first coordinate point, the second coordinate point, the third coordinate point, and the fourth coordinate point, respectively;

[0106] For practical applications, please refer to Figure 3 As shown, the obtained contour angle.

[0107] The reference point setting module is configured with a reference point setting strategy, which includes:

[0108] Connect the first coordinate point and the second coordinate point to obtain a line segment, which is marked as the first line segment; connect the second coordinate point and the third coordinate point to obtain a line segment, which is marked as the second line segment;

[0109] Draw a perpendicular line from the midpoint of the first line segment to the first line segment, and mark it as the first perpendicular line; draw a perpendicular line from the midpoint of the second line segment to the second line segment, and mark it as the second perpendicular line.

[0110] Obtain the intersection of the first and second perpendicular lines and mark it as the reference point for the detection dots. The method for obtaining the reference point for the detection dots is a method of obtaining the center of a circle. Since the dots of the light guide plate are circular, the reference point for the detection dots is the center of the dots. At the same time, the points of defective parts are excluded by the reference difference, making the obtained center of the circle more accurate and making the subsequently obtained dot detection values ​​more valuable for comparison.

[0111] For practical applications, please refer to Figure 4 As shown, the obtained detection network reference points.

[0112] The detection value acquisition module is used to acquire the detection value of the detection point based on the detection point outline and the detection point reference point;

[0113] The detection value acquisition module is configured with a detection value acquisition strategy, which includes:

[0114] Mark the pixels on the detected dot outline as detected outline pixels;

[0115] Obtain the distance from the reference point of the detected halftone dot to the pixel of the detected contour, and mark it as the halftone dot detection value; the halftone dot detection value is the radius of the halftone dot. If most of the halftone dot detection values ​​are too large, it may be due to the temperature and pressure settings being too high. If most of the halftone dot detection values ​​are too small, it may be due to the temperature and pressure settings being too low. If a small portion of the halftone dot detection values ​​are too large or too small, it may be due to the generation of incomplete halftone dots.

[0116] In practical applications, for example, the detection value of a single mesh point is 0.041 mm.

[0117] The threshold acquisition module is used to acquire a first qualified threshold and a second qualified threshold based on the image of the first normal light guide plate after hot pressing.

[0118] The threshold acquisition module is configured with a threshold acquisition strategy, which includes:

[0119] The image after the first normal number of light guide plates are hot-pressed is marked as a normal image; the detection values ​​of all the dots in the outline of each detection dot in the normal image are obtained and marked as normal distances;

[0120] Obtain the range of normal distances, labeled C1 to Cn; divide the range of fall comparison values ​​evenly into d parts, labeled as normal parts; the length of each normal part is (Cn-C1)÷d; the starting and ending values ​​of each normal part are C1+[(Cn-C1)÷d]×e and C1+[(Cn-C1)÷d]×(e+1), respectively; where e is a positive integer from 0 to d-1; the normal parts are divided to observe the distribution of normal distances, for example, a is set to 10;

[0121] In practical applications, the normal distance is obtained from 0.038mm to 0.048mm. The gray values ​​from 0 to 255 are evenly divided into 10 parts. The length of the 10 normal parts is (0.048-0.038)÷10=0.001. When e is 0, the starting value and ending value of each normal part are 0.038+[(0.048-0.038)÷8]×0=0.038 and 0.038+[(0.048-0.038)÷8]×(1+1)=0.039, respectively. Therefore, a normal part is 0.038 to 0.039.

[0122] Count the number of each normal part and mark it as the number of normal parts;

[0123] Sort the number of normal parts in ascending order of their starting values ​​to obtain normal sequence numbers, which are positive integers starting from 1;

[0124] Obtain the frequency of normal distances and label it Z;

[0125] The threshold for the number of normal cases is calculated as: G = f × (Z ÷ d); where Z is the fall classification threshold, f is the ratio from 0 to 1; f is set to obtain the normal part with a smaller number of normal cases, so f is set to a small value, for example, f is 0.1;

[0126] In practical applications, the frequency of normal distance is 1212. The threshold for the number of normal distances is calculated as: G = 0.1 × (1212 ÷ 10) = 12. The result is rounded to the nearest integer. Therefore, the threshold for the number of normal distances is 12.

[0127] The system checks whether the number of normal parts is less than the normal number threshold according to the normal sequence number from smallest to largest. If not, the check stops. If so, the number of normal parts is deleted, and the system continues to check whether the number of normal parts is less than the normal number threshold until the number of normal parts is greater than or equal to the normal number threshold. The system then obtains the starting value of the normal part corresponding to the number of normal parts on the leftmost side and marks it as the first qualified threshold.

[0128] The system checks if the number of normal parts is less than a threshold number, starting from the largest normal sequence number. If not, the check stops; otherwise, the normal part number is deleted, and the system continues checking if the next normal part number is less than the threshold number, until the normal part number is greater than or equal to the threshold number. The system then retrieves the end value of the normal part corresponding to the rightmost normal part number and marks it as the second acceptable threshold. This method can delete abnormally large or small normal distances, obtaining a more precise range of normal distances.

[0129] In practical applications, the number of normal portions is sorted from smallest to largest according to their starting values ​​as 4, 21, ..., 32, 10. 4 and 10, which are less than the threshold of 12, are deleted. The starting value of the normal portion corresponding to the leftmost 21 after deletion is 0.39mm to 0.40mm, which is the first acceptable threshold. The ending value of the normal portion corresponding to the leftmost 32 after deletion is 0.46mm to 0.47mm, which is the second acceptable threshold.

[0130] The judgment module is used to determine whether the light guide plate is qualified after hot pressing based on the dot detection value, the first qualified threshold and the second qualified threshold;

[0131] The judgment module is configured with judgment strategies, which include:

[0132] Determine if all dot detection values ​​are greater than or equal to the first acceptable threshold and less than or equal to the second acceptable threshold. If yes, generate a signal indicating that the light guide plate heat pressing is qualified; otherwise, generate a signal indicating that the light guide plate heat pressing is unqualified. If the dot detection value is less than the first acceptable threshold, it may be due to the temperature and pressure settings being too low or the dots being incomplete. If the dot detection value is greater than the second acceptable threshold, it may be due to the temperature and pressure settings being too high or the dots being incomplete. Therefore, determine whether the light guide plate is qualified.

[0133] In practical applications, for example, if the detection value of a dot is 0.041mm, which satisfies the condition that 0.041mm is greater than the first qualified threshold of 0.39mm and 0.041mm is less than 0.47mm, then it is a normal distance. If both conditions are met, then a signal indicating that the heat pressing of the light guide plate is qualified is generated.

[0134] Example 2, please refer toFigure 5 As shown, an AOI inspection method for a light guide plate after hot pressing includes the following steps:

[0135] Step S1: Obtain the image of the light guide plate after hot pressing and mark it as the hot-pressed image.

[0136] Step S2 involves performing grayscale processing on the hot-pressed image to obtain a grayscale image after hot pressing; Step S2 includes the following sub-steps:

[0137] Step S201: Obtain the R value, G value, and B value of any pixel in the hot-pressed image, and label them as the detected R value, detected G value, and detected B value, respectively; calculate the mean of the detected R value, detected G value, and detected B value, and label it as the detected gray value; replace the detected R value, detected G value, and detected B value with the detected gray value, and represent each pixel in the hot-pressed image with the corresponding detected gray value to obtain the hot-pressed grayscale image.

[0138] Step S3 involves binarizing the hot-pressed grayscale image to obtain a binarized image after hot pressing. Step S3 includes the following sub-steps:

[0139] Step S301: Divide the gray values ​​from 0 to 255 into a equal parts, and mark them as equal parts; the length of each equal part is 255÷a; the starting value and ending value of each equal part are (255÷a)×b and (255÷a)×(b+1), respectively; where b is a positive integer from 0 to a-1.

[0140] Step S302: Obtain the number of detected grayscale values ​​in each evenly divided part and mark it as the number of detections;

[0141] Step S303: Draw a histogram with grayscale value as the X-axis, number of detections as the Y-axis, and evenly divided parts as histogram intervals, and mark it as the detection grayscale histogram;

[0142] Step S304: Mark the evenly divided portion of the detection grayscale histogram that is greater than the number of adjacent detections on the left and right sides as the first target portion;

[0143] Step S305: Obtain the first target part with the smallest initial value and mark it as the second target part;

[0144] Step S306: Obtain the first target part with the largest initial value and mark it as the third target part;

[0145] Step S307: Obtain the evenly divided part with the smallest initial value between the second target part and the third target part, and mark it as the fourth target part;

[0146] Step S308: Calculate the mean of the starting and ending values ​​of the fourth target part and mark it as the dot binarization threshold;

[0147] Step S309: In the grayscale image after hot pressing, the detected grayscale values ​​that are greater than the halftone dot binarization threshold are set to 0, and the detected grayscale values ​​that are less than or equal to the binarization threshold are set to 255, so as to obtain the binarized image after hot pressing.

[0148] Step S4: Obtain the outline of the detected dots based on the binarized image after hot pressing; Step S4 includes the following sub-steps:

[0149] Step S401: In the binarized image after hot pressing, obtain the pixel with a detected gray value of 0 that is adjacent to the detected gray value of 255 and mark it as the initial pixel; determine whether there are pixels with a detected gray value of 0 around the initial pixel. If not, obtain a new initial pixel; if so, mark the initial pixel as the edge starting pixel.

[0150] Step S402: Starting with the pixel adjacent to the left of the edge starting pixel, and using the edge starting pixel as the rotation center, rotate clockwise to search for the first pixel whose grayscale value changes from 255 to 0, and mark it as the edge search pixel; then use the edge search pixel as the rotation center to obtain a new edge search pixel, and stop the operation when the new edge search pixel is the same as the edge starting pixel, and obtain the contour composed of the edge starting pixel and all the edge search pixels, and mark it as the detection dot contour.

[0151] Step S5: Obtain reference points for the detected dots based on the dot contour; Step S5 includes the following sub-steps:

[0152] Step S501: Obtain the outlines of all detection dots in the binarized image after hot pressing;

[0153] Step S502: Draw a second number of coordinate points at equal intervals on the detected dot outline and mark them as outline coordinate points;

[0154] Step S503: Connect the contour coordinate points sequentially along the contour of the detected dots to obtain line segments, and mark them as contour connection line segments;

[0155] Step S504: Obtain the included angle between all adjacent contour connecting line segments and mark it as the contour angle;

[0156] Step S505: Calculate the mean value of the contour angles and mark it as the mean value of the contour angles;

[0157] Step S506: Calculate the absolute value of the difference between all contour angles and the mean of contour angles, and mark it as the reference difference;

[0158] Step S507: Sort the reference differences from smallest to largest, obtain the intersection points of the contour angles corresponding to the first four reference differences, and mark them as the first coordinate point, the second coordinate point, the third coordinate point, and the fourth coordinate point, respectively.

[0159] Step S508: Connect the first coordinate point and the second coordinate point to obtain a line segment, which is marked as the first line segment; connect the second coordinate point and the third coordinate point to obtain a line segment, which is marked as the second line segment;

[0160] Step S509: Draw a perpendicular line from the midpoint of the first line segment to the first line segment, and mark it as the first perpendicular line; draw a perpendicular line from the midpoint of the second line segment to the second line segment, and mark it as the second perpendicular line.

[0161] Step S510: Obtain the intersection of the first perpendicular line and the second perpendicular line, and mark it as the reference point for the detection network.

[0162] Step S6: Obtain the detection value of the halftone dots based on the detected halftone dot outline and the detected halftone dot reference point; Step S6 includes the following sub-steps:

[0163] Step S601: Mark the pixels on the detected dot outline as detected outline pixels;

[0164] Step S602: Obtain the distance from the reference point of the detected halftone dot to the pixel point of the detected contour, and mark it as the halftone dot detection value.

[0165] Step S7: Obtain a first qualified threshold and a second qualified threshold based on the image after the first number of normal light guide plates are hot-pressed; Step S7 includes the following sub-steps:

[0166] Step S701: Mark the image after the first normal number of light guide plates are hot-pressed as a normal image; obtain all the halftone detection values ​​of each halftone dot contour in the normal image and mark them as normal distances;

[0167] Step S702: Obtain the range of normal distances, marked as C1 to Cn; divide the range of fall comparison values ​​evenly into d parts, marked as normal parts; the length of each normal part is (Cn-C1)÷d; the starting value and ending value of each normal part are C1+[(Cn-C1)÷d]×e and C1+[(Cn-C1)÷d]×(e+1), respectively; where e is a positive integer from 0 to d-1.

[0168] Step S703: Count the number of each normal part and mark it as the number of normal parts;

[0169] Step S704: Sort the number of normal parts in ascending order according to the starting value of the normal parts to obtain the normal sequence number. The normal sequence number is a positive integer starting from 1.

[0170] Step S705: Obtain the frequency of normal distances and label it as Z;

[0171] Step S706, calculate the normal number threshold as: G=f×(Z÷d;where Z is the fall classification threshold, and f is the ratio from 0 to 1;

[0172] Step S707: Determine whether the number of normal parts is less than the normal number threshold according to the normal sequence number from smallest to largest. If not, stop the judgment directly. If yes, delete the number of normal parts and continue to determine whether the number of normal parts is less than the normal number threshold until the number of normal parts is greater than or equal to the normal number threshold. Obtain the starting value of the normal part corresponding to the number of normal parts on the leftmost side at this time and mark it as the first qualified threshold.

[0173] Step S708: Determine whether the number of normal parts is less than the normal number threshold according to the normal sequence number from largest to smallest. If not, stop the judgment directly. If yes, delete the number of normal parts and continue to determine whether the number of normal parts is less than the normal number threshold until the number of normal parts is greater than or equal to the normal number threshold. Obtain the end value of the normal part corresponding to the number of normal parts on the far right at this time and mark it as the second qualified threshold.

[0174] Step S8: Determine whether the light guide plate is qualified after hot pressing based on the dot detection value, the first qualified threshold, and the second qualified threshold; Step S8 includes the following sub-steps:

[0175] Step S801: Determine whether the detection values ​​of all dots are greater than or equal to the first qualified threshold and less than or equal to the second qualified threshold. If yes, generate a signal that the heat pressing of the light guide plate is qualified; otherwise, generate a signal that the heat pressing of the light guide plate is unqualified.

[0176] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0177] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings or direct couplings or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

Claims

1. A light guide plate hot line after pressing AOI detection system, characterized in that, The method comprises the following steps: An image acquisition module, a grayscale processing module, a binaryzation module, a contour extraction module, a reference point setting module, a detection value acquisition module, a threshold value acquisition module, and a judgment module; The image acquisition module is configured to acquire an image of the light guide plate after heat pressing, which is marked as a heat-pressed image; The grayscale processing module is configured to perform grayscale processing on the heat-pressed image to obtain a heat-pressed grayscale image; The binaryzation module is configured to perform binaryzation processing on the heat-pressed grayscale image to obtain a heat-pressed binaryzation image; The contour extraction module is configured to acquire a detection dot contour based on the heat-pressed binaryzation image; The reference point setting module is configured to acquire a detection dot reference point based on the detection dot contour; The detection value acquisition module is configured to acquire a dot detection value based on the detection dot contour and the detection dot reference point; The threshold value acquisition module is configured to acquire a first qualified threshold value and a second qualified threshold value based on a first number of normal light guide plates after heat pressing; The judgment module is configured to judge whether the light guide plate after heat pressing is qualified based on the dot detection value, the first qualified threshold value, and the second qualified threshold value; The reference point setting module is configured with a reference point setting strategy, which comprises the following steps: acquiring all detection dot contours in the heat-pressed binaryzation image; drawing a second number of coordinate points at every same interval on the detection dot contour, which are marked as contour coordinate points; connecting the contour coordinate points along the detection dot contour to obtain line segments, which are marked as contour connection line segments; acquiring the included angles between all adjacent contour connection line segments, which are marked as contour included angles; calculating the mean value of the contour included angles, which is marked as a contour included angle mean value; calculating the absolute values of the differences between all contour included angles and the contour included angle mean value, which are marked as reference difference values; sorting the reference difference values from small to large, and acquiring the intersection points of the contour included angles corresponding to the first four reference difference values, which are respectively marked as a first coordinate point, a second coordinate point, a third coordinate point, and a fourth coordinate point; connecting the first coordinate point and the second coordinate point to obtain a line segment, which is marked as a first line segment; connecting the second coordinate point and the third coordinate point to obtain a line segment, which is marked as a second line segment; drawing a perpendicular line of the first line segment through the midpoint of the first line segment, which is marked as a first perpendicular line; drawing a perpendicular line of the second line segment through the midpoint of the second line segment, which is marked as a second perpendicular line; acquiring the intersection point of the first perpendicular line and the second perpendicular line, which is marked as a detection dot reference point; The detection value acquisition module is configured with a detection value acquisition strategy, which comprises the following steps: marking the pixel points on the detection dot contour as detection contour pixel points; acquiring the distance from the detection dot reference point to the detection contour pixel points, which is marked as a dot detection value.

2. The system according to claim 1, wherein, The grayscale processing module is configured with a grayscale processing strategy, which comprises the following steps: acquiring the R value, the G value, and the B value of any one pixel point in the heat-pressed image, which are respectively marked as a detection R value, a detection G value, and a detection B value; calculating the mean value of the detection R value, the detection G value, and the detection B value, which is marked as a detection grayscale value; replacing the detection R value, the detection G value, and the detection B value with the detection grayscale value, so that each pixel point in the heat-pressed image is represented by the corresponding detection grayscale value, thereby obtaining the heat-pressed grayscale image.

3. The system according to claim 2, wherein, The binarization module is configured with a binarization strategy, and the binarization strategy comprises: Divide the gray values from 0 to 255 into a parts evenly, marked as evenly divided parts; the length of each evenly divided part is 255÷a; the start value and the end value of each evenly divided part are (255÷a)×b and (255÷a)×(b+1) respectively; wherein b is a positive integer from 0 to a-1; Obtain the number of detected gray values in each evenly divided part, marked as the number of detections; Draw a histogram with the gray values as the X axis, the number of detections as the Y axis, and the evenly divided parts as the histogram intervals, marked as the detected gray histogram; Mark the evenly divided part with the number of detections greater than the adjacent number of detections on the left and right sides in the detected gray histogram as the first target part; Obtain the first target part with the minimum start value, marked as the second target part; Obtain the first target part with the maximum start value, marked as the third target part; Obtain the evenly divided part with the minimum start value between the second target part and the third target part, marked as the fourth target part; Calculate the average of the start value and the end value of the fourth target part, marked as the dot binarization threshold value; In the gray image after hot pressing, set the detected gray values greater than the dot binarization threshold value to 0, and set the detected gray values less than or equal to the binarization threshold value to 255, to obtain the binarized image after hot pressing.

4. The system according to claim 3, wherein, The contour extraction is configured with a contour extraction strategy, and the contour extraction strategy comprises: In the binarized image after hot pressing, obtain the pixel point with the detected gray value of 0 adjacent to the detected gray value of 255, marked as the initial pixel point; judge whether the initial pixel point contains the pixel point with the detected gray value of 0 around the initial pixel point, if not, obtain a new initial pixel point; if yes, mark the initial pixel point as the edge start pixel point; Take the pixel point adjacent to the left of the edge start pixel point as the start, and take the edge start pixel point as the rotation center, rotate in the clockwise direction to search for the first pixel point with the detected gray value changing from 255 to 0, marked as the edge search pixel point; then take the edge search pixel point as the rotation center to obtain a new edge search pixel point, until the new edge search pixel point is the same as the edge start pixel point, stop the operation, and obtain the contour composed of the edge start pixel point and all the edge search pixel points, marked as the detected dot contour.

5. The system according to claim 4, wherein the system further comprises a light source and a camera. The threshold value acquisition module is configured with a threshold value acquisition strategy, and the threshold value acquisition strategy comprises: Mark the image of the first number of normal light guide plates after hot pressing as a normal image; obtain all the dot detection values of each detected dot contour in the normal image, marked as the normal distance; Obtain the range of the normal distance, marked as C1 to Cn; divide the range of the normal distance into d parts evenly, marked as normal parts; the length of each normal part is (Cn-C1)÷d; the start value and the end value of each normal part are C1+[(Cn-C1)÷d]×e and C1+[(Cn-C1)÷d]×(e+1) respectively; wherein e is a positive integer from 0 to d-1; Count the number of each normal part, marked as the number of normal parts; Sort the normal part numbers in ascending order according to the starting values of the normal parts to obtain normal sequence numbers, which are positive integers starting from 1; Obtain the frequency of normal distances, denoted as Z; Calculate the normal number threshold value as: G = f x (Z ÷ d); wherein G is the normal number threshold value, and f is a proportional value ranging from 0 to 1; Determine whether the normal part number is less than the normal number threshold value in ascending order of the normal sequence numbers, if not, stop the determination directly, if yes, delete the normal part number, continue to determine whether the next normal part number is less than the normal number threshold value until the normal part number is greater than or equal to the normal number threshold value, and stop the determination; obtain the starting value of the normal part corresponding to the leftmost normal part number at this time, denoted as the first qualified threshold value; Determine whether the normal part number is less than the normal number threshold value in descending order of the normal sequence numbers, if not, stop the determination directly, if yes, delete the normal part number, continue to determine whether the next normal part number is less than the normal number threshold value until the normal part number is greater than or equal to the normal number threshold value, and stop the determination; obtain the end value of the normal part corresponding to the rightmost normal part number at this time, denoted as the second qualified threshold value.

6. The system according to claim 5, wherein the system further comprises a light source and a camera. The judgment module is configured with a judgment strategy, and the judgment strategy comprises: Determine whether all the net point detection values are greater than or equal to the first qualified threshold value and less than or equal to the second qualified threshold value, if yes, generate a signal that the light guide plate hot pressing is qualified, if not, generate a signal that the light guide plate hot pressing is unqualified.

7. A light guide plate hot line after AOI detection method, suitable for the light guide plate hot line after AOI detection system of any one of claims 1-6, characterized in that, The method comprises the following steps: Obtain the image after the light guide plate is hot pressed, denoted as the hot-pressed image; Perform gray processing on the hot-pressed image to obtain a hot-pressed gray image; Perform binaryzation processing on the hot-pressed gray image to obtain a hot-pressed binaryzation image; Obtain the detection net point contour based on the hot-pressed binaryzation image; Obtain the detection net point reference point based on the net point contour; Obtain the net point detection value based on the detection net point contour and the detection net point reference point; Obtain the first qualified threshold value and the second qualified threshold value based on the image after the first number of normal light guide plates are hot pressed; Determine whether the light guide plate after hot pressing is qualified based on the net point detection value, the first qualified threshold value and the second qualified threshold value.

Images