A detection method and system for a liquid crystal panel

Abstract

The present invention relates to a method and system for detecting a liquid crystal panel. The detection method is used to implement edge detection and / or defect detection of the liquid crystal panel. Among them, the edge detection includes the following steps: obtaining a measurement result image of the liquid crystal panel to be measured; extracting the region of interest of the measurement result image; obtaining the corner coordinate points and / or edge contours in the region of interest of the measurement result image; measuring the grinding accuracy of the liquid crystal panel to be measured according to the edge contour to obtain a grinding accuracy detection result; and / or, measuring the cutting accuracy of the liquid crystal panel to be measured according to the edge contour to obtain a cutting accuracy detection result; and / or, measuring the chamfering accuracy of the liquid crystal panel to be measured according to the corner coordinate points and the edge contour to obtain a chamfering accuracy detection result; outputting and storing the grinding accuracy detection result and / or the cutting accuracy detection result and / or the chamfering accuracy detection result; The present invention improves the efficiency of edge detection and defect detection and the stability of the detection results.

Description

Technical Field

[0001] The present invention relates to the technical field of liquid crystal panel detection, and particularly to a detection method and system for liquid crystal panels. Background Art

[0002] At present, the detection of liquid crystal panels includes two parts: edge measurement and defect detection. It mainly relies on manual work. The corresponding size measurement is carried out on the screen edge, and then the defect judgment is made by human eyes. Edge measurement mainly includes three parts: grinding measurement, cutting measurement and chamfer measurement. The manual measurement method currently adopted in factories is mainly carried out under a microscope. For grinding measurement, multiple points are taken at designated positions or random positions in the grinding area of the screen, and then the width of the grinding area is measured; for cutting measurement, the distances from the screen edge to the upper and lower mark points are measured; for chamfer measurement, the chamfers at the upper left corner, lower left corner, upper right corner and lower right corner of the screen are measured. Its detection is time-consuming and laborious, and the accuracy of the detection results of manual detection is related to the experience of the detection personnel, which affects the accuracy of the final detection results.

[0003] Defect detection mainly detects four types of defects: notch, breakage, crack and protrusion. The defect detection method currently adopted in factories is mainly to observe with human eyes to find out whether these four types of defects exist on the screen in turn. This way of manual measurement and defect detection is time-consuming and laborious, with low efficiency, a cumbersome process and strong subjectivity, and is affected by factors such as personnel emotions, eyesight, environmental light, and inconsistent judgment criteria of different human eyes. Summary of the Invention

[0004] In order to solve the technical problems in the prior art that the stability and uniformity of the detection results of the edge measurement and defect detection of liquid crystal panels by manual work are poor, the present invention provides a detection method and system for liquid crystal panels.

[0005] The technical solution of the present invention to solve the above technical problems is as follows:

[0006] A detection method for liquid crystal panels, the detection method is used to realize the edge detection and / or defect detection of liquid crystal panels; wherein, the edge detection includes the following steps:

[0007] Obtain the measurement result image of the liquid crystal panel to be measured;

[0008] Extract the region of interest of the measurement result image;

[0009] Obtain the corner coordinate points and / or edge contours in the region of interest of the measurement result image;

[0010] Measure the grinding accuracy of the liquid crystal panel to be measured according to the edge profile to obtain a grinding accuracy detection result; and / or measure the cutting accuracy of the liquid crystal panel to be measured according to the edge profile to obtain a cutting accuracy detection result; and / or measure the chamfering accuracy of the liquid crystal panel to be measured according to the angular coordinate points and the edge profile to obtain a chamfering accuracy detection result;

[0011] Output and store the grinding accuracy detection result and / or the cutting accuracy detection result and / or the chamfering accuracy detection result.

[0012] The beneficial effect of the present invention is that by determining the edges and angular coordinate points in the measurement result image of the liquid crystal panel to be measured, parameters such as the edge dimensions of the liquid crystal panel to be measured are measured, which is convenient for improving the efficiency of edge detection and the stability of the detection result.

[0013] On the basis of the above technical solution, the present invention can also be improved as follows.

[0014] Further, the defect detection includes the following steps:

[0015] Extract the edge defect image of the liquid crystal panel to be measured in the measurement result image;

[0016] Classify the edge defect image of the liquid crystal panel to be measured by comparing the edge defect image of the liquid crystal panel to be measured with a defect comparison model; wherein, the defect comparison model is a computer vision recognition model;

[0017] Output and store the classified edge defect image of the liquid crystal panel to be measured.

[0018] Further, classify the edge defect image of the liquid crystal panel to be measured by comparing the edge defect image of the liquid crystal panel to be measured with a defect comparison model; specifically, it includes the following steps:

[0019] Establish a computer vision model;

[0020] Use edge defect sample data of various types of liquid crystal panels to train the computer vision model to obtain multiple defect comparison models;

[0021] Compare the edge defect image of the liquid crystal panel to be measured with multiple defect comparison models respectively to obtain multiple edge defect images of the liquid crystal panel to be measured corresponding to the defect comparison models;

[0022] Classify all the edge defect images of the liquid crystal panel to be measured.

[0023] The beneficial effects of adopting the above further solution are as follows: By comparing the edge defect image of the liquid crystal panel to be measured with the defect comparison model to identify the preset defects in the model, automatic identification of edge defects is achieved through computer vision recognition method, which improves the recognition efficiency. At the same time, the recognition standard is unified, and the stability of the detection results of the liquid crystal panel is improved.

[0024] Further, storing the edge detection result and the edge defect image of the liquid crystal panel to be measured that has been classified specifically includes the following steps:

[0025] Store the edge detection result;

[0026] Store the edge defect images of the liquid crystal panel to be measured that have been classified separately by category.

[0027] Further, extracting the region of interest of the measurement result image specifically includes the following steps:

[0028] Obtain the boundary of the liquid crystal panel to be measured in the measurement result image;

[0029] By screening the boundary of the liquid crystal panel to be measured in the measurement result image, determine the outer frame of the liquid crystal panel to be measured in the measurement result image;

[0030] Determine the region of interest according to the outer frame of the liquid crystal panel to be measured;

[0031] Alternatively, extracting the region of interest of the measurement result image specifically includes the following steps:

[0032] Extract the background region of the measurement result image;

[0033] Use the preset background region area, preset background region width value, and preset background region height value to perform region screening on the measurement result image to determine the region of interest.

[0034] The beneficial effects of adopting the above further solution are as follows: By screening the boundary of the liquid crystal panel to be measured to determine the region of interest of the measurement result image, the detection versatility of the liquid crystal panel can be improved in the case of different sizes of liquid crystal panels; by using the preset background region area, preset background region width value, and preset background region height value to perform region screening on the measurement result image, the same type of liquid crystal panels produced in batches can be uniformly processed, improving the detection efficiency.

[0035] Further, measuring the grinding accuracy of the liquid crystal panel to be measured according to the edge contour of the liquid crystal panel to be measured to obtain a grinding accuracy detection result, specifically including the following steps:

[0036] Based on the threshold segmentation method, obtain the binary image of the measurement result image;

[0037] Based on the straight line fitting method, obtain the boundary lines on both sides of the grinding area of the liquid crystal panel to be measured in the binary image of the measurement result image according to the edge contour of the liquid crystal panel to be measured;

[0038] Calculate the distance between the boundary lines on both sides of the grinding area of the liquid crystal panel to be measured to obtain the grinding width value;

[0039] Calculate the area between the boundary lines on both sides of the grinding area of the liquid crystal panel to be measured to obtain the grinding area value;

[0040] Calculate the rectangularity of the rectangular area corresponding to the boundary lines on both sides of the grinding area of the liquid crystal panel to be measured to obtain the grinding rectangularity value;

[0041] Calculate the height of the area between the boundary lines on both sides of the grinding area of the liquid crystal panel to be measured to obtain the grinding height value;

[0042] Wherein, the grinding precision detection result includes the grinding width value, the grinding area value, the grinding rectangularity value and the grinding height value.

[0043] The beneficial effect of adopting the above further solution is that by using the straight line fitting method to obtain the boundary lines on both sides of the grinding area of the liquid crystal panel to be measured, the amount of data for detection calculation can be increased, and the detection result can be closer to the actual boundary line, improving the detection precision.

[0044] Further, calculating the distance between the boundary lines on both sides of the grinding area of the liquid crystal panel to be measured to obtain the grinding width value specifically includes the following steps:

[0045] Segment the grinding area of the liquid crystal panel to be measured in the binary image to obtain a plurality of segmented grinding areas;

[0046] Calculate the width values of the boundary lines on both sides of the grinding area in each segmented grinding area respectively;

[0047] Take the average of the width values of the boundary lines on both sides of the grinding area in all segmented grinding areas to obtain the grinding width value.

[0048] The beneficial effect of adopting the above further solution is that the accuracy of the detection result can be improved by the method of segmented detection and averaging.

[0049] Further, measure the cutting precision of the liquid crystal panel to be measured according to the edge contour of the liquid crystal panel to be measured to obtain the cutting precision detection result, including the following steps:

[0050] When there are upper and lower cross Mark points on the measurement result diagram, connect the upper and lower cross Mark points, calculate the distance from any point on the edge contour to the connection line, and obtain the cutting accuracy value; when there is no cross Mark point or only one cross Mark point on the measurement result diagram, determine any target point on the edge contour, identify the corresponding marking position of any target point on the edge contour in the measurement result diagram, calculate the distance from any target point on the edge contour to the corresponding marking position, and obtain the cutting accuracy value;

[0051] Among them, the cutting accuracy detection result includes the cutting accuracy value.

[0052] Furthermore, according to the corner coordinate points and the edge contour of the liquid crystal panel to be measured, measure the chamfer accuracy of the liquid crystal panel to be measured, and obtain the chamfer accuracy detection result, including the following steps:

[0053] Use the adaptive threshold algorithm to obtain the chamfer feature marking positions in the region of interest of the measurement result image, and determine the chamfer detection region;

[0054] Segment and extract the chamfer detection region to obtain the chamfer region image;

[0055] According to the preset chamfer area feature value, preset chamfer width feature value, and preset chamfer height feature value, screen out the final chamfer image in the chamfer region image;

[0056] Determine the minimum circumscribed rectangle of the final chamfer according to the corner coordinate points and the edge contour;

[0057] According to the minimum circumscribed rectangle of the final chamfer, calculate the final chamfer width value and the final chamfer height value of the final chamfer;

[0058] Among them, the chamfer accuracy detection result includes the final chamfer width value and the final chamfer height value

[0059] To solve the above technical problems, the present invention also provides the following technical solutions:

[0060] A detection system for a liquid crystal panel, the detection system is used for edge detection and / or defect detection of the liquid crystal panel; among them, the edge detection system includes,

[0061] A data acquisition module, configured to acquire a measurement result image of the liquid crystal panel to be measured;

[0062] A data processing module, configured to extract the region of interest of the measurement result image, and obtain the corner coordinate points and / or edge contour in the measurement result image;

[0063] An accuracy detection module, configured to measure the grinding accuracy of the liquid crystal panel to be measured according to the edge profile, and obtain a grinding accuracy detection result; and / or, configured to measure the cutting accuracy of the liquid crystal panel to be measured according to the edge profile, and obtain a cutting accuracy detection result; and / or, configured to measure the chamfering accuracy of the liquid crystal panel to be measured according to the angular coordinate points and the edge profile, and obtain a chamfering accuracy detection result;

[0064] An output module, configured to output the grinding accuracy detection result and / or the cutting accuracy detection result and / or the chamfering accuracy detection result;

[0065] A storage module, configured to store the grinding accuracy detection result and / or the cutting accuracy detection result and / or the chamfering accuracy detection result. Description of the Drawings

[0066] Figure 1 It is a flowchart of a method for detecting a liquid crystal panel in an embodiment of the present invention. Detailed Embodiments

[0067] The principles and features of the present invention will be described below with reference to the accompanying drawings. The examples cited are only used to explain the present invention and are not intended to limit the scope of the present invention.

[0068] Embodiment 1

[0069] As Figure 1 shown, this embodiment provides a method for detecting a liquid crystal panel, and the detection method is used to implement edge detection and / or defect detection of the liquid crystal panel; wherein, the edge detection includes the following steps:

[0070] S1. Obtain a measurement result image of the liquid crystal panel to be measured;

[0071] S2. Extract the region of interest of the measurement result image; the specific steps are: obtain the boundary of the liquid crystal panel to be measured in the measurement result image;

[0072] By screening the boundary of the liquid crystal panel to be measured in the measurement result image, determine the outer frame of the liquid crystal panel to be measured in the measurement result image;

[0073] Determine the region of interest according to the outer frame of the liquid crystal panel to be measured;

[0074] Or, the specific steps are:

[0075] Use the preset background region area, the preset background region width value, and the preset background region height value to perform region screening on the measurement result image to determine the region of interest. If the region of interest is inaccurate, the problem can be solved by increasing the threshold parameter.

[0076] S3. Obtain the angular coordinate points and / or the edge profile in the region of interest of the measurement result image.

[0077] S4. Measure the grinding accuracy of the liquid crystal panel to be measured according to the edge profile to obtain a grinding accuracy detection result; and / or, measure the cutting accuracy of the liquid crystal panel to be measured according to the edge profile to obtain a cutting accuracy detection result; and / or, measure the chamfering accuracy of the liquid crystal panel to be measured according to the angular coordinate points and the edge profile to obtain a chamfering accuracy detection result; specifically, measuring the grinding accuracy of the liquid crystal panel to be measured according to the edge profile of the liquid crystal panel to be measured to obtain a grinding accuracy detection result includes the following steps:

[0078] Based on the threshold segmentation method, obtain the binary image of the measurement result image;

[0079] Based on the straight line fitting method, obtain the two side boundary lines of the grinding area of the liquid crystal panel to be measured in the binary image of the measurement result image;

[0080] Calculate the distance between the two side boundary lines of the grinding area of the liquid crystal panel to be measured to obtain the grinding width value;

[0081] Calculate the area between the two side boundary lines of the grinding area of the liquid crystal panel to be measured to obtain the grinding area value;

[0082] Calculate the rectangularity of the rectangular area corresponding to the two side boundary lines of the grinding area of the liquid crystal panel to be measured to obtain the grinding rectangularity value;

[0083] Calculate the height of the area between the two side boundary lines of the grinding area of the liquid crystal panel to be measured to obtain the grinding height value.

[0084] Calculating the distance between the two side boundary lines of the grinding area of the liquid crystal panel to be measured to obtain the grinding width value includes the following steps:

[0085] Segment the grinding area of the liquid crystal panel to be measured in the binary image to obtain a plurality of segmented grinding areas;

[0086] Calculate the width values of the two side boundary lines of the grinding area in each segmented grinding area respectively;

[0087] Take the average of the width values of the two side boundary lines of the grinding area in all the segmented grinding areas to obtain the grinding width value.

[0088] Wherein, the grinding accuracy detection result includes the grinding width value, the grinding area value, the grinding rectangularity value and the grinding height value.

[0089] Measuring the cutting accuracy of the liquid crystal panel to be measured according to the edge profile of the liquid crystal panel to be measured to obtain a cutting accuracy detection result includes the following steps:

[0090] When there are upper and lower cross Mark points on the measurement result graph, connect the upper and lower cross Mark points, calculate the distance from any point on the edge contour to the connection line, and obtain the cutting accuracy value; when there is no cross Mark point or only one cross Mark point on the measurement result graph, determine any target point on the edge contour, identify the corresponding mark position of any target point on the edge contour in the measurement result graph, calculate the distance from any target point on the edge contour to the corresponding mark position, and obtain the cutting accuracy value. Among them, the cutting accuracy detection result includes the cutting accuracy value.

[0091] Measure the chamfering accuracy of the liquid crystal panel to be measured according to the corner coordinate points and edge contour of the liquid crystal panel to be measured, including the following steps:

[0092] Using an adaptive threshold algorithm, obtain the chamfering feature flag bits in the region of interest of the measurement result image, and determine the chamfering detection region;

[0093] Segment and extract the chamfering detection region to obtain a chamfering region image;

[0094] According to the preset chamfering area feature value, preset chamfering width feature value, and preset chamfering height feature value, screen out the final chamfering image in the chamfering region image;

[0095] Determine the minimum circumscribed rectangle of the final chamfering according to the corner coordinate points and the edge contour;

[0096] According to the minimum circumscribed rectangle of the final chamfering, calculate the final chamfering width value and final chamfering height value of the final chamfering;

[0097] Among them, the chamfering accuracy detection result includes the final chamfering width value and the final chamfering height value.

[0098] S5. Output and store the grinding accuracy detection result and / or cutting accuracy detection result and / or chamfering accuracy detection result.

[0099] Embodiment 2

[0100] Based on Embodiment 1, this embodiment provides a detection method for a liquid crystal panel, and the detection method is used to implement edge detection and / or defect detection of the liquid crystal panel; among them, the defect detection includes the following steps:

[0101] S10. Extract the edge defect image of the liquid crystal panel to be measured in the measurement result image;

[0102] S11. Classify the edge defect image of the liquid crystal panel to be measured by comparing the edge defect image of the liquid crystal panel to be measured with a defect comparison model; among them, the defect comparison model is a computer vision recognition model; specifically includes the following steps:

[0103] Build a computer vision model;

[0104] Use various types of liquid crystal panel edge defect sample data to train the computer vision model to obtain multiple defect comparison models; among them, the various types of liquid crystal panel edge defect sample data at least include notch sample data, breakage sample data, crack sample data, and protrusion sample data;

[0105] Compare the edge defect images of the liquid crystal panel to be measured with multiple defect comparison models respectively to obtain multiple edge defect images of the liquid crystal panel to be measured corresponding to the defect comparison models;

[0106] Classify all the edge defect images of the liquid crystal panel to be measured.

[0107] Among them, the method for obtaining the edge defect image of the liquid crystal panel to be measured is the adaptive threshold method; the idea of the adaptive threshold is not to calculate the threshold of the global image, but to calculate its local threshold according to the brightness distribution of different regions of the image. Therefore, for different regions of the image, different thresholds can be calculated adaptively, so it is called the adaptive threshold method. It is applicable to the defect segmentation of images with uneven gray levels; the algorithm assumes that the image pixels can be divided into background and target parts according to the threshold. Then, calculate the optimal threshold to distinguish these two types of pixels, so that the discrimination degree of the two types of pixels is the largest. The specific calculation formula is as follows:

[0108]

[0109]

[0110] N1 + N2 = M × N

[0111] ω1 + ω2 = 1

[0112] μ = μ1 × ω1 + μ2 × ω2

[0113] g = ω1 × (μ - μ1) 2 + ω2 × (μ - μ2) 2

[0114] N1 represents the number of pixels whose pixel gray level is greater than the segmentation threshold of the target and the background; N2 represents the number of pixels whose pixel gray level is less than the segmentation threshold of the target and the background; M × N represents the size of the image; ω1 represents the proportion of the number of target pixels in the whole image; ω2 represents the proportion of the number of background pixels in the whole image; μ1 represents the average gray level of the target pixels; μ2 represents the average gray level of the background pixels; μ represents the total average gray level of the image; g represents the between-class variance.

[0115] S12. Output the classified edge defect images of the liquid crystal panel to be measured;.

[0116] S13. Store the edge defect images of the liquid crystal panel to be measured that have been classified; specifically, it includes the following steps: Store the edge defect images of the liquid crystal panel to be measured that have been classified according to their categories respectively.

[0117] In this embodiment, by comparing the edge defect images of the liquid crystal panel to be measured with the defect comparison model, the preset defects in the model are identified, realizing the automatic identification of edge defects through the computer vision recognition method, improving the recognition efficiency, unifying the recognition standard at the same time, and improving the stability of the detection results of the liquid crystal panel.

[0118] Embodiment 3

[0119] Based on Embodiment 1, this embodiment provides a detection system for a liquid crystal panel, including a data acquisition module, a data processing module, a precision detection module, a defect detection module, an output module, and a storage module.

[0120] The data acquisition module is used to acquire the measurement result image of the liquid crystal panel to be measured;

[0121] The data processing module is used to extract the region of interest of the measurement result image and obtain the corner coordinate points and / or edge contours of the liquid crystal panel to be measured in the measurement result image;

[0122] The precision detection module is used to measure the grinding precision of the liquid crystal panel to be measured according to the edge contour of the liquid crystal panel to be measured, and obtain a grinding precision detection result; and / or, used to measure the cutting precision of the liquid crystal panel to be measured according to the edge contour of the liquid crystal panel to be measured, and obtain a cutting precision detection result; and / or, used to measure the chamfering precision of the liquid crystal panel to be measured according to the corner coordinate points and edge contour of the liquid crystal panel to be measured, and obtain a chamfering precision detection result;

[0123] The defect detection module is used to extract the edge defect images of the liquid crystal panel to be measured in the measurement result image; classify the edge defect images of the liquid crystal panel to be measured by comparing the edge defect images of the liquid crystal panel to be measured with a defect comparison model; wherein, the defect comparison model is a computer vision recognition model;

[0124] The output module is used to output the grinding precision detection result and / or the cutting precision detection result and / or the chamfering precision detection result and / or the edge defect images of the liquid crystal panel to be measured that have been classified;

[0125] The storage module is used to store the grinding precision detection result and / or the cutting precision detection result and / or the chamfering precision detection result and / or the edge defect images of the liquid crystal panel to be measured that have been classified.

[0126] Specifically, the above-mentioned data acquisition module, data processing module, accuracy detection module, defect detection module, and output module are all computer programs or collections of functional codes in a computer program; the storage module can be a computer's memory, internal memory, flash memory, and external storage devices, etc.

[0127] In the embodiment of the present invention, by using a computer program to obtain a measurement result image of a liquid crystal panel to be measured, and determining the edge and corner coordinate points in the measurement result image of the liquid crystal panel to be measured, parameters such as the edge size of the liquid crystal panel to be measured are measured. Compared with manual detection, the efficiency of edge detection and the stability of the detection result are improved; by comparing the edge defect image of the liquid crystal panel to be measured with a defect comparison model to identify the preset defects in the model, automatic identification of edge defects is achieved through computer vision recognition method, improving the recognition efficiency, unifying the recognition standard at the same time, and improving the stability of the detection result of the liquid crystal panel.

[0128] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the concept and principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A detection method for a liquid crystal panel, characterized in that The described detection method is used to implement edge detection and / or defect detection of a liquid crystal panel; wherein, the edge detection includes the following steps: Obtain a measurement result image of the liquid crystal panel to be measured; Extract the region of interest of the measurement result image; Obtain the corner coordinate points and / or edge contours in the region of interest of the measurement result image; Measure the grinding accuracy of the liquid crystal panel to be measured according to the edge contour to obtain a grinding accuracy detection result; and / or, measure the cutting accuracy of the liquid crystal panel to be measured according to the edge contour to obtain a cutting accuracy detection result; and / or, measure the chamfering accuracy of the liquid crystal panel to be measured according to the corner coordinate points and the edge contour to obtain a chamfering accuracy detection result; Output and store the grinding accuracy detection result and / or cutting accuracy detection result and / or chamfering accuracy detection result; Measuring the grinding accuracy of the liquid crystal panel to be measured according to the edge contour to obtain a grinding accuracy detection result specifically includes the following steps: Based on the threshold segmentation method, obtain a binary image of the measurement result image; Based on the straight line fitting method, obtain the boundary lines on both sides of the grinding area of the liquid crystal panel to be measured in the binary image of the measurement result image according to the edge contour of the liquid crystal panel to be measured; Calculate the distance between the boundary lines on both sides of the grinding area of the liquid crystal panel to be measured to obtain a grinding width value; Calculate the area between the boundary lines on both sides of the grinding area of the liquid crystal panel to be measured to obtain a grinding area value; Calculate the rectangularity of the rectangular area corresponding to the boundary lines on both sides of the grinding area of the liquid crystal panel to be measured to obtain a grinding rectangularity value; Calculate the height of the area between the boundary lines on both sides of the grinding area of the liquid crystal panel to be measured to obtain a grinding height value; Wherein, the grinding accuracy detection result includes the grinding width value, the grinding area value, the grinding rectangularity value and the grinding height value.

2. The detection method of the liquid crystal panel according to claim 1, wherein The defect detection includes the following steps: Extract the edge defect image of the liquid crystal panel to be measured in the measurement result image; Classify the edge defect image of the liquid crystal panel to be measured by comparing the edge defect image of the liquid crystal panel to be measured with a defect comparison model; wherein, the defect comparison model is a computer vision recognition model; Output and store the classified edge defect image of the liquid crystal panel to be measured.

3. The detection method of the liquid crystal panel according to claim 2, wherein, Classifying the edge defect image of the liquid crystal panel to be measured by comparing the edge defect image of the liquid crystal panel to be measured with a defect comparison model specifically includes the following steps: Establish a computer vision model; Use edge defect sample data of various types of liquid crystal panels to train the computer vision model to obtain multiple defect comparison models; Compare the edge defect image of the liquid crystal panel to be measured with multiple defect comparison models respectively to obtain multiple edge defect images of the liquid crystal panel to be measured corresponding to the defect comparison models; Classify all the edge defect images of the liquid crystal panel to be measured.

4. The inspection method of the liquid crystal panel according to claim 3, characterized in that Storing the classified edge defect image of the liquid crystal panel to be measured specifically includes the following steps: Store the classified edge defect images of the liquid crystal panel to be measured separately by category.

5. The detection method of the liquid crystal panel according to claim 1, wherein, Extracting the region of interest of the measurement result image specifically includes the following steps: Obtain the boundary of the liquid crystal panel to be measured in the measurement result image; Determine the outer frame of the liquid crystal panel to be measured in the measurement result image by screening the boundary of the liquid crystal panel to be measured in the measurement result image; Determine the region of interest according to the outer frame of the liquid crystal panel to be measured; Alternatively, extract the region of interest of the measurement result image, which specifically includes the following steps: Extract the background region of the measurement result image; Perform region screening on the measurement result image by using a preset background region area, a preset background region width value, and a preset background region height value to determine the region of interest.

6. The detection method of the liquid crystal panel according to claim 1, wherein Calculate the distance between the boundary lines on both sides of the grinding area of the liquid crystal panel to be measured to obtain the grinding width value, which specifically includes the following steps: Segment the grinding area of the liquid crystal panel to be measured in the binary image to obtain a plurality of segmented grinding regions; Calculate the width values of the boundary lines on both sides of the grinding area in each segmented grinding region respectively; Take the average of the width values of the boundary lines on both sides of the grinding area in all the segmented grinding regions to obtain the grinding width value.

7. The detection method of the liquid crystal panel according to claim 1 or 5, characterized in that, Measure the cutting accuracy of the liquid crystal panel to be measured according to the edge contour to obtain a cutting accuracy detection result, which specifically includes the following steps: When there are two cross Mark points, upper and lower, on the measurement result graph, connect the upper and lower cross Mark points, and calculate the distance from any point on the edge contour to the connection line to obtain the cutting accuracy value; when there is no cross Mark point or only one cross Mark point on the measurement result graph, determine any target point on the edge contour, identify the corresponding mark position of any target point on the edge contour in the measurement result graph, and calculate the distance from any target point on the edge contour to the corresponding mark position to obtain the cutting accuracy value; Wherein, the cutting accuracy detection result includes the cutting accuracy value.

8. The detection method of the liquid crystal panel according to claim 1 or 5, characterized in that, Measure the chamfering accuracy of the liquid crystal panel to be measured according to the angular coordinate points and the edge contour to obtain a chamfering accuracy detection result, which specifically includes the following steps: Use an adaptive threshold algorithm to obtain the chamfer feature flag positions in the region of interest of the measurement result image and determine the chamfer detection region; Perform segmentation extraction on the chamfer detection region to obtain a chamfer region image; According to a preset chamfer area feature value, a preset chamfer width feature value, and a preset chamfer height feature value, screen out the final chamfer image in the chamfer region image; Determine the minimum circumscribed rectangle of the final chamfer according to the angular coordinate points and the edge contour; According to the minimum circumscribed rectangle of the final chamfer, calculate the final chamfer width value and the final chamfer height value of the final chamfer; Wherein, the chamfering accuracy detection result includes the final chamfer width value and the final chamfer height value.

9. A detection system for a liquid crystal panel, characterized in that The detection system is used for edge detection and / or defect detection of a liquid crystal panel; wherein, the detection system includes, A data acquisition module for acquiring a measurement result image of a liquid crystal panel to be measured; A data processing module for extracting the region of interest of the measurement result image and acquiring the angular coordinate points and / or edge contour in the measurement result image; The precision detection module is used to measure the grinding precision of the liquid crystal panel to be measured according to the edge contour, and obtain a grinding precision detection result; and / or, to measure the cutting precision of the liquid crystal panel to be measured according to the edge contour, and obtain a cutting precision detection result; and / or, to measure the chamfering precision of the liquid crystal panel to be measured according to the angular coordinate points and the edge contour, and obtain a chamfering precision detection result; The output module is used to output the grinding precision detection result and / or the cutting precision detection result and / or the chamfering precision detection result; The storage module is used to store the grinding precision detection result and / or the cutting precision detection result and / or the chamfering precision detection result; Measuring the grinding precision of the liquid crystal panel to be measured according to the edge contour, and obtaining a grinding precision detection result, specifically including the following steps: Based on the threshold segmentation method, obtain the binary image of the measurement result image; Based on the straight line fitting method, obtain the two side boundary lines of the grinding area of the measurement result image of the liquid crystal panel to be measured according to the edge contour of the liquid crystal panel to be measured; Calculate the distance between the two side boundary lines of the grinding area of the liquid crystal panel to be measured to obtain a grinding width value; Calculate the area between the two side boundary lines of the grinding area of the liquid crystal panel to be measured to obtain a grinding area value; Calculate the rectangularity of the rectangular area corresponding to the two side boundary lines of the grinding area of the liquid crystal panel to be measured to obtain a grinding rectangularity value; Calculate the height of the area between the two side boundary lines of the grinding area of the liquid crystal panel to be measured to obtain a grinding height value; Wherein, the grinding precision detection result includes the grinding width value, the grinding area value, the grinding rectangularity value and the grinding height value.

Images