Display panel aggregation defect detection method and detection device

Abstract

The present invention provides a detection method and a detection device for clustering defects of a display panel, the detection method comprising: obtaining a defect point distribution image of a current display panel; performing DBSCAN clustering processing on the defect point distribution image to determine whether there is a clustering area; dividing the clustering area into at least a first number of sub-areas, obtaining a clustering parameter corresponding to each sub-area according to the defect points in each sub-area, determining whether the clustering parameter of each sub-area is greater than a first preset value and storing the defect point distribution image; recording a sub-area whose clustering parameter is greater than the first preset value as a current area of ​​interest, comparing the defect point features of the current area of ​​interest with the defect point features of past areas of interest to obtain similar parameters of the current area of ​​interest, determining whether the similar parameters are greater than a second preset value; and reporting the occurrence of similar clustering defects. The present invention can improve the detection accuracy while realizing the induction and arrangement of clustering defect types.

Description

Technical Field

[0001] The present invention relates to the field of testing, and in particular to a method and a device for detecting aggregation defects of a display panel. Background Art

[0002] During the production process of large display panels, multiple processes are required. After each process is completed, the entire large panel needs to be inspected. If the inspection is performed manually, due to the low efficiency of manual inspection, more than half of the panels cannot be inspected in time, and there is a risk that defective panels will flow to subsequent processes. With the development of science and technology, the degree of automated production is getting higher and higher. AOI (Automated Optical Inspection) is a commonly used method for defect detection of display panels, but in actual applications, there are still problems in using AOI for clustered defect detection. For example, when the machine operating parameters are not adjusted in time according to the current product, the AOI test results may show abnormalities, but in fact the display panel is normal. In addition, since the defects that appear in the display panel during the production process are unexpected and intermittent, even if the same type of clustered defects occur in different time periods, because there is generally no systematic sorting, it is necessary to rely on the memory and acuity of the production personnel to determine whether there are similar clustered defects before. The risk of omission is high and the search and confirmation takes a long time, so it is impossible to achieve a systematic sorting of the same type of clustered defects, which is not conducive to statistical management of data in the later stage.

[0003] Therefore, how to improve the detection accuracy while achieving the induction and organization of clustered defect types is an urgent problem to be solved. Summary of the invention

[0004] The present invention provides a method and a device for detecting aggregation defects of a display panel to solve the above technical problems.

[0005] In order to achieve the above object, the present invention provides a method for detecting aggregation defects of a display panel, the detection method comprising:

[0006] Step A: obtaining a defect point distribution image of the current display panel, wherein defect points of at least one type of defect are distributed on the defect point distribution image;

[0007] Step B: performing DBSCAN clustering processing on the defect point distribution image, and judging whether there is a clustering area for the at least one type of defect on the defect point distribution image according to a preset clustering condition, and if so, executing step C;

[0008] Step C: at least dividing the clustering area into a first number of sub-areas, obtaining a clustering parameter corresponding to each sub-area according to the defect points of the at least one type of defect in each sub-area, determining whether the clustering parameter of each sub-area is greater than a first preset value, and if so, storing a defect point distribution image and executing step D;

[0009] Step D: Record the position of the sub-region whose aggregation parameter is greater than the first preset value and define it as the current region of interest, compare the defect point features of the current region of interest with the defect point features of the past region of interest corresponding to the defect point distribution image stored in the past, obtain the similarity parameter of the current region of interest, and determine whether the similarity parameter is greater than the second preset value. If so, execute step E;

[0010] Step E: Report the occurrence of similar clustered defects.

[0011] As an optional technical solution, the clustering parameter includes a first parameter and a second parameter, the first parameter is the number of defect points in each sub-region, and the second parameter is the ratio of the number of defect points in each sub-region to the area of ​​the corresponding sub-region.

[0012] As an optional technical solution, the first preset value includes a first threshold and a second threshold. If the first parameter of a sub-area is greater than the first threshold and the second parameter is greater than the second threshold, the aggregation parameter of the sub-area is greater than the first preset value, and it is determined that there is an aggregation defect in the defect point distribution image.

[0013] As an optional technical solution, step C includes drawing a corresponding thermal map based on the defect point distribution image, and the thermal map includes a second number of sub-areas, and the second number is not less than the first number.

[0014] As an optional technical solution, the defect point distribution image includes multiple first dividing lines and multiple second dividing lines, the multiple first dividing lines extend along a first direction, the multiple second dividing lines extend along a second direction, the first direction is not parallel to the second direction, the multiple first dividing lines and the multiple second dividing lines define multiple cells, each sub-area includes at least one cell, and the area of ​​interest includes at least one sub-area.

[0015] As an optional technical solution, the defect point feature includes a first feature, a second feature and a third feature, wherein the first feature is the ratio of the number of defect points clustered in the region of interest to the total number of defect points in the defect point distribution image; the second feature is the number of defect points in the region of interest; and the third feature is the ratio of the number of cells with defect points in the region of interest to the total number of cells in the region of interest.

[0016] As an optional technical solution, the second preset value includes a third threshold, a fourth threshold and a fifth threshold. If the first feature is greater than the third threshold, the second feature is greater than the fourth threshold and the third feature is greater than the fifth threshold, then it is confirmed that the similarity parameter is greater than the second preset value, and it is determined that a similar aggregation defect occurs.

[0017] As an optional technical solution, step A includes that the at least one type of defects includes at least a first type of defects on the current display panel, and the first type of defects is the largest defect of the currently produced product counted based on current production data.

[0018] As an optional technical solution, the detection method also includes step F, after a preset time interval, updating the first major defect of the product currently being produced according to current production data, and returning to step A.

[0019] In addition, the present invention also proposes a detection device for clustered defects of a display panel, which includes an acquisition module, a processing module and a prompt module. The acquisition module is used to acquire a defect point distribution image of the current display panel, on which defect points of at least one type of defect are distributed; the processing module is used to process the defect point distribution image to obtain a processing result, and the processing module is electrically connected to the acquisition module; the prompt module is used to acquire and report the processing result, and the prompt module is electrically connected to the processing module, wherein the processing module performs DBSCAN clustering processing on the defect point distribution image, and determines whether there is a clustering area for the at least one type of defect on the defect point distribution image according to a preset clustering condition, and if so, at least divides the clustering area into a first number Measure sub-regions, obtain the aggregation parameter corresponding to each sub-region according to the defect points of at least one type of defects in each sub-region, judge whether the aggregation parameter of each sub-region is greater than a first preset value, if so, store the defect point distribution image, record the position of the sub-region where the aggregation parameter is greater than the first preset value and define it as the current region of interest, compare the defect point features of the current region of interest with the defect point features of the past region of interest corresponding to the defect point distribution pattern stored in the past, obtain the similarity parameter of the current region of interest, judge whether the similarity parameter is greater than a second preset value, if so, transmit the processing result to the prompt module.

[0020] The present invention provides a detection method and a detection device for clustered defects of a display panel. First, a defect point distribution image of the current display panel is obtained and then DBSCAN clustering processing is performed to preliminarily determine whether there is a clustered area in the defect point distribution image. Then, the clustered area is divided into at least a first number of sub-areas. The clustering parameters corresponding to each sub-area are obtained according to the defect points in each sub-area to determine whether it is a true cluster, thereby improving the accuracy of detecting clustered defects. Furthermore, the defect point distribution image with clustered defects is stored, and the corresponding sub-area position is recorded as the current area of ​​interest. The defect point features of the current area of ​​interest are compared with the defect point features of the past area of ​​interest to determine whether there are similar clustered defects, thereby achieving the induction and organization of the types of clustered defects, so as to analyze and improve the manufacturing environment, manufacturing machines, manufacturing parameters, etc. of the display panel, and improve the yield rate of the display panel. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The specific implementation modes of the present invention are further described in detail below in conjunction with the accompanying drawings.

[0022] Figure 1 It is a flow chart of the method for detecting aggregation defects of a display panel of the present invention;

[0023] Figure 2 is a schematic diagram of a defect point distribution image in the present invention;

[0024] Figure 3 is a schematic diagram of defect point distribution images of multiple display panels in the present invention;

[0025] Figure 4 It is a schematic diagram of a device for detecting aggregation defects of a display panel according to the present invention. DETAILED DESCRIPTION

[0026] In order to more clearly illustrate the present invention, the present invention is further described below in conjunction with preferred embodiments and accompanying drawings. Similar components in the accompanying drawings are represented by the same reference numerals. It should be understood by those skilled in the art that the content specifically described below is illustrative rather than restrictive, and should not be used to limit the scope of protection of the present invention.

[0027] See also Figure 1 , Figure 1The flowchart of the detection method of the display panel aggregation defect of the present invention is shown. The present invention provides a detection method of the display panel aggregation defect, the defect point distribution image of the current display panel is clustered, and it is preliminarily determined whether there is an aggregation defect in the defect point distribution image. If so, the next step of judgment is performed, at least the aggregation area in the defect point distribution image is divided into a first number of sub-areas, and the aggregation parameters corresponding to each sub-area are obtained according to the defect points in each sub-area to determine whether it is true aggregation. In one embodiment, true aggregation refers to severe aggregation; further, the defect point distribution image with aggregation defects is stored, the position of the sub-area with aggregation defects is recorded and defined as the current area of ​​interest, the defect point features of the current area of ​​interest are compared with the defect point features of the past area of ​​interest, and the similar parameters of the current area of ​​interest are obtained to determine whether there are similar aggregation defects. If so, the occurrence of similar aggregation defects is reported, so that the same type of aggregation defects can be archived and retained, which is convenient for subsequent management personnel to analyze and process the data of aggregation defects in the defect point distribution image.

[0028] like Figure 1 As shown, the method for detecting aggregation defects of a display panel comprises the following steps:

[0029] Step A (S110): Obtain a defect point distribution image of the current display panel, on which defect points of at least one type of defect are distributed; in one embodiment, the defect point distribution image can be formed by performing image capture and comparison processing on the current display panel through AOI equipment.

[0030] Step B (S120): performing DBSCAN clustering processing on the defect point distribution image, and determining whether there is a clustering area for at least one type of defect on the defect point distribution image according to a preset clustering condition, and if so, executing step C (S130);

[0031] In this embodiment, the DBSCAN clustering algorithm is a density-based spatial clustering algorithm. The management personnel can set two parameters, the neighborhood radius distance and the minimum number of points in the neighborhood, as needed. The DBSCAN clustering algorithm is used to perform density division on the clusters of the defect point distribution image and mark the areas with higher density with colors.

[0032] Step C (S130): Divide the clustering area into at least a first number of sub-areas, obtain a clustering parameter corresponding to each sub-area according to the defect points of the at least one type of defect in each sub-area, determine whether the clustering parameter of each sub-area is greater than a first preset value, and if so, store the defect point distribution image and execute step D (S140);

[0033] In one embodiment, the aggregation parameter includes a first parameter and a second parameter, the first parameter is the number of defect points in each sub-region, the second parameter is the ratio of the number of defect points in each sub-region to the area of ​​the corresponding sub-region, and the first preset value includes a first threshold and a second threshold, the first threshold corresponds to the first parameter, and the second threshold corresponds to the second parameter; only when the first parameter of a sub-region is greater than the first threshold and the second parameter is greater than the second threshold, will it be confirmed that the aggregation parameter of the sub-region is greater than the first preset value, thereby determining that there is an aggregation defect in the defect point distribution image.

[0034] For example, the first threshold is pre-set to 30 and the second threshold is pre-set to 0.2%. Then, if the number of defect points in a sub-region in the defect point distribution image of the current display panel is greater than 30 and the ratio of the number of defect points to the area of ​​the corresponding sub-region reaches 0.2%, it is considered that there are clustered defects in the defect point distribution image; or, the first threshold is pre-set to 10 and the second threshold is pre-set to 10%. Then, if the number of defect points in a sub-region in the defect point distribution image of the current display panel is greater than 10 and the ratio of the number of defect points to the area of ​​the corresponding sub-region reaches 10%, it can be considered that there are indeed clustered defects in the defect point distribution image. In actual applications, the management personnel can set the first threshold and the second threshold in the first preset value according to actual needs, but are not limited to this.

[0035] In this embodiment, DBSCAN clustering processing is performed in step B to preliminarily confirm whether there is a clustering phenomenon (or a clustering area) in the defect point distribution image; then in step C, further judgment is made on the defect point distribution image with the clustering area to confirm whether the clustering phenomenon of the current display panel has reached a serious level. If it is judged to be serious (that is, the clustering parameter is confirmed to be greater than the first preset value in step C, which means that there are indeed serious clustering defects in the defect point distribution image), the detection results can be reported to the management personnel and the abnormal data can be recorded and stored. The management personnel can monitor the corresponding defect point distribution image according to the notification results; if it is judged to be not serious (that is, the clustering parameter obtained according to the defect points in the defect point distribution image in step C does not reach the first preset value), in other words, the first parameter and the second parameter of each sub-area are not greater than the corresponding first threshold and second threshold at the same time, then it is judged that the clustering parameter does not reach the first preset value, and the defect point distribution image can be filtered. In practical applications, the management personnel can be notified by sending emails, messages, etc., but it is not limited to this.

[0036] Step D (S140): recording the position of the sub-region where the aggregation parameter is greater than the first preset value and defining it as the current region of interest, comparing the defect point features of the current region of interest with the defect point features of the past region of interest corresponding to the defect point distribution image stored in the past, obtaining a similarity parameter of the current region of interest, and determining whether the similarity parameter is greater than the second preset value. If so, executing step E (S150);

[0037] Step E (S150): reporting the occurrence of similar cluster defects.

[0038] In one embodiment, the manager can select the region of interest in the defect point distribution image of the current display panel with clustered defects to compare with the defect point features of the past region of interest corresponding to the defect point distribution image stored in the past. After the defect point distribution image of the current display panel is analyzed, the defect point distribution image is stored as one of the updated past regions of interest. If, in the subsequent inspection process, the region of interest of the defect point distribution image of another display panel has similar defects to the region of interest in the previously stored defect point distribution image, the same type of defect point distribution images can be summarized and sorted out to analyze and improve the manufacturing environment, manufacturing machine, manufacturing parameters, etc. of the display panel to improve the yield rate of the display panel.

[0039] In one embodiment, at least one type of defect is distributed in the defect point distribution image in step A, and the at least one type of defect may include at least the first type of defect on the current display panel, and the first type of defect is the largest defect of the currently produced product counted according to the current production data (i.e., the most defects generated by the currently produced product counted in combination with the current production data), and the detection method further includes step F (S160), after a preset time interval, updating the largest defect of the currently produced product according to the current production data, and returning to step A (S110). In another embodiment, the first type of defect may include the first two defects on the current display panel, but is not limited thereto.

[0040] In one embodiment, the preset interval time is 15 minutes, so that the first major defect of the currently produced product can be updated according to the current production data every 15 minutes, and whether the current display panel has a cluster defect can be determined according to the updated first major defect (or first type of defect). In actual operation, the preset interval time is not limited to this.

[0041] In one embodiment, step C includes drawing a corresponding heat map according to the defect point distribution image, and the heat map includes a second number of sub-areas, and the second number is not less than the first number. Figure 2 , Figure 2 FIG. 1 is a schematic diagram of a defect point distribution image in the present invention. Figure 2 As shown, in combination with step C and step D, the entire defect point distribution image is divided, and the defect point distribution image includes multiple first cutting lines 10 and multiple second cutting lines 20, multiple first cutting lines 10 extend along the first direction D1, multiple second cutting lines 20 extend along the second direction D2, the first direction D1 is not parallel to the second direction D2 (in this embodiment, the first direction D1 is perpendicular to the second direction D2, but not limited to this), multiple first cutting lines 10 and multiple second cutting lines 20 define multiple cells 30, each sub-area 40 includes at least one cell 30, and the region of interest 50 includes at least one sub-area 40. Figure 2 As shown, if the density of defect points in the cell 30 is large, the color of the corresponding cell 30 is darker. Conversely, if the density of defect points in the cell 30 is small, the color of the corresponding cell 30 is lighter. The judge (such as a manager or a machine, etc.) can intuitively determine from the thermal map whether there is a clustered area in the defect point distribution image.

[0042] See also Figure 3 , Figure 3 FIG. 1 is a schematic diagram showing defect point distribution images of multiple display panels in the present invention. Figure 3 As shown, in this embodiment, the defect point distribution image includes 6 sub-areas 40, one sub-area 40 includes 9 cells 30, and the defect point distribution image includes 54 cells 30 in total. If a certain type of defect is detected to have regional clustering in the upper left corner of a defect point distribution image 1, the manager sets the sub-area 40 in the upper left corner as the area of ​​interest 50, that is, the area of ​​interest 50 only includes one sub-area 40 at this time, stores the defect point distribution image 1 and records the defect point features of the area of ​​interest 50 as the past area of ​​interest corresponding to the defect point distribution image stored in the past. In the subsequent detection process, if it is found that the upper left corner of another defect point distribution image 2 (which can be considered as the current area of ​​interest of the defect point distribution image of the current display panel) has the phenomenon of clustering of at least one type of defect occurrence area mentioned above, it can be further confirmed whether there are past areas of interest with similar clustered defects in the past areas of interest judged by the defect point distribution images stored in the past. If so, if the similarity parameter is greater than the second preset value at this time, the defect point features of the area of ​​interest 50 of the defect point distribution image 1 stored in the past and the defect point distribution image 2 of the current display panel can be centrally analyzed and processed so as to improve the manufacturing environment / machine parameters of the display panel, etc. and improve the yield rate.

[0043] In actual applications, managers can divide the defect point distribution image as needed. The more sub-areas 40 the defect point distribution image contains and the more cells 30 a sub-area 40 contains, the higher the detection accuracy of clustered defects. The division of sub-areas and cells can be performed according to actual needs, not limited to the foregoing.

[0044] In one embodiment, the defect point feature includes a first feature, a second feature and a third feature, wherein the first feature is the ratio of the number of defect points clustered in the region of interest to the total number of defect points in the defect point distribution image; the second feature is the number of defect points in the region of interest; and the third feature is the ratio of the number of cells with defect points in the region of interest to the total number of cells in the region of interest. The second preset value includes a third threshold, a fourth threshold and a fifth threshold, wherein the third threshold corresponds to the first feature, the fourth threshold corresponds to the second feature, and the fifth threshold corresponds to the third feature. Only when the first feature is greater than the third threshold, the second feature is greater than the fourth threshold and the third feature is greater than the fifth threshold, will it be confirmed that the similarity parameter is greater than the second preset value, thereby determining that a similar cluster defect occurs. In other words, as long as at least one of the first feature, the second feature and the third feature is not greater than the third threshold, the fourth threshold and the fifth threshold in the corresponding second preset value, it is determined that the similarity parameter does not reach the second preset value.

[0045] For example, according to the defect point distribution image 1 of the previous display panel stored in the past, it is judged that there is a clustered defect distribution, and the past region of interest is recorded, and the third threshold in the second preset value can be set to 90%, the fourth threshold to 110, and the fifth threshold to 75%. After obtaining the defect point distribution image 2 of the current display panel, if the number of defect points clustered in the region of interest 50 in the defect point distribution image 2 is 217, the total number of defect points in the defect point distribution image is 235, and the number of cells with defect points in the region of interest is 7, the first feature in the defect point distribution image 2 is calculated to be 92%, the second feature is 217, and the third feature is 77%, then the first feature, the second feature and the third feature in the defect point distribution image 2 are all greater than the corresponding third threshold, fourth threshold and fifth threshold in the second preset value. At this time, it is judged that the defect point distribution image 2 of the current display panel and the defect point distribution image of the display panel stored in the past have similar clustering, the data of the defect point distribution image 2 is archived and retained, and the occurrence of similar clustering defects is notified.

[0046] In practical applications, the manager may also select multiple adjacent sub-areas as the regions of interest as needed, for example, select two sub-areas as the regions of interest, but the present invention is not limited thereto.

[0047] See also Figure 4 , Figure 4 FIG. 2 is a schematic diagram of a device for detecting aggregation defects of a display panel according to the present invention. Figure 4As shown, the detection device 100 for clustered defects of a display panel of the present invention comprises an acquisition module 110, a processing module 120 and a prompt module 130. The acquisition module 110 is used to acquire a defect point distribution image of the current display panel, on which defect points of at least one type of defect are distributed; the processing module 120 is used to process the defect point distribution image to obtain a processing result, and the processing module 120 is electrically connected to the acquisition module 110; the prompt module 130 is used to acquire and report the processing result, and the prompt module 130 is electrically connected to the processing module 120, wherein the processing module 120 performs DBSCAN clustering processing on the defect point distribution image, and determines whether there is a clustering area for at least one type of defect on the defect point distribution image according to a preset clustering condition, and if so, at least the clustering area Divide into a first number of sub-regions, obtain a clustering parameter corresponding to each sub-region according to defect points of at least one type of defects in each sub-region, determine whether the clustering parameter of each sub-region is greater than a first preset value, if so, store the defect point distribution image, record the position of the sub-region where the clustering parameter is greater than the first preset value and define it as the current region of interest, compare the defect point features of the current region of interest with the defect point features of the past region of interest corresponding to the defect point distribution pattern stored in the past, obtain similarity parameters of the current region of interest, determine whether the similarity parameter is greater than a second preset value, if so, transmit the processing result to the prompt module 130.

[0048] In summary, the present invention provides a detection method and device for clustered defects of display panels. First, the defect point distribution image of the current display panel is obtained and then DBSCAN clustering processing is performed to preliminarily determine whether there is a clustered area in the defect point distribution image. Then, the clustered area is divided into at least a first number of sub-areas. The clustering parameters corresponding to each sub-area are obtained according to the defect points in each sub-area to determine whether it is a true cluster, thereby improving the accuracy of detecting clustered defects. Furthermore, the defect point distribution image with clustered defects is stored, and the corresponding sub-area position is recorded as the current area of ​​interest. The defect point features of the current area of ​​interest are compared with the defect point features of the past area of ​​interest to determine whether there are similar clustered defects, thereby achieving the induction and organization of the types of clustered defects, so as to analyze and improve the manufacturing environment, manufacturing machines, manufacturing parameters, etc. of the display panel, and improve the yield of the display panel.

[0049] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not limitations on the implementation methods of the present invention. For ordinary technicians in the relevant field, other different forms of changes or modifications can be made based on the above description. It is impossible to list all the implementation methods here. All obvious changes or modifications derived from the technical solution of the present invention are still within the protection scope of the present invention.

Claims

1. A method for detecting aggregation defects of a display panel, characterized in that: The test method includes: Step A: obtaining a defect point distribution image of the current display panel, wherein defect points of at least one type of defect are distributed on the defect point distribution image; Step B: performing DBSCAN clustering processing on the defect point distribution image, and judging whether there is a clustering area for the at least one type of defect on the defect point distribution image according to a preset clustering condition, and if so, executing step C; Step C: at least dividing the clustering area into a first number of sub-areas, obtaining a clustering parameter corresponding to each sub-area according to the defect points of the at least one type of defect in each sub-area, determining whether the clustering parameter of each sub-area is greater than a first preset value, and if so, storing a defect point distribution image and executing step D; Step D: Record the position of the sub-region whose aggregation parameter is greater than the first preset value and define it as the current region of interest, compare the defect point features of the current region of interest with the defect point features of the past region of interest corresponding to the defect point distribution image stored in the past, obtain the similarity parameter of the current region of interest, and determine whether the similarity parameter is greater than the second preset value. If so, execute step E; Step E: Report similar clustered defects; The clustering parameter includes a first parameter and a second parameter, the first parameter is the number of defect points in each sub-region, and the second parameter is the ratio of the number of defect points in each sub-region to the area of ​​the corresponding sub-region.

2. The detection method according to claim 1, characterized in that: The first preset value includes a first threshold and a second threshold. If the first parameter of a sub-area is greater than the first threshold and the second parameter is greater than the second threshold, the aggregation parameter of the sub-area is greater than the first preset value, and it is determined that there is an aggregation defect in the defect point distribution image.

3. The detection method according to claim 1, characterized in that: Step C includes drawing a corresponding thermal map according to the defect point distribution image, and the thermal map includes a second number of sub-areas, and the second number is not less than the first number.

4. The detection method according to claim 3, characterized in that: The defect point distribution image includes multiple first dividing lines and multiple second dividing lines, the multiple first dividing lines extend along a first direction, the multiple second dividing lines extend along a second direction, the first direction is not parallel to the second direction, the multiple first dividing lines and the multiple second dividing lines define multiple cells, each sub-area includes at least one cell, and the area of ​​interest includes at least one sub-area.

5. The detection method according to claim 4, characterized in that: The defect point feature includes a first feature, a second feature and a third feature, wherein the first feature is the ratio of the number of defect points clustered in the region of interest to the total number of defect points in the defect point distribution image; the second feature is the number of defect points in the region of interest; and the third feature is the ratio of the number of cells with defect points in the region of interest to the total number of cells in the region of interest.

6. The detection method according to claim 5, characterized in that: The second preset value includes a third threshold, a fourth threshold and a fifth threshold. If the first feature is greater than the third threshold, the second feature is greater than the fourth threshold and the third feature is greater than the fifth threshold, it is confirmed that the similarity parameter is greater than the second preset value, and it is determined that a similar aggregation defect occurs.

7. The detection method according to claim 1, characterized in that: Step A includes that the at least one type of defect includes at least a first type of defect on the current display panel, and the first type of defect is the largest defect of the currently produced product counted according to current production data.

8. The detection method according to claim 7, characterized in that: The detection method also includes step F, after a preset time interval, updating the first major defect of the product currently being produced according to current production data, and returning to step A.

9. A device for detecting aggregation defects of a display panel, characterized in that: The detection device comprises: An acquisition module, used for acquiring a defect point distribution image of the current display panel, where defect points of at least one type of defect are distributed on the defect point distribution image; A processing module, used for processing the defect point distribution image to obtain a processing result, the processing module is electrically connected to the acquisition module; as well as a prompt module, used to obtain and notify the processing result, the prompt module is electrically connected to the processing module, Wherein, the processing module performs DBSCAN clustering processing on the defect point distribution image, determines whether there is a clustering area for the at least one type of defect on the defect point distribution image according to a preset clustering condition, and if so, divides the clustering area into at least a first number of sub-areas, obtains a clustering parameter corresponding to each sub-area according to the defect points of the at least one type of defect in each sub-area, and determines whether the clustering parameter of each sub-area is greater than a first preset value, and if so, stores the defect point distribution image, records the position of the sub-area whose clustering parameter is greater than the first preset value and defines it as the current region of interest, compares the defect point features of the current region of interest with the defect point features of the past region of interest corresponding to the defect point distribution pattern stored in the past, obtains a similarity parameter of the current region of interest, and determines whether the similarity parameter is greater than a second preset value, and if so, transmits the processing result to the prompt module; The clustering parameter includes a first parameter and a second parameter, the first parameter is the number of defect points in each sub-region, and the second parameter is the ratio of the number of defect points in each sub-region to the area of ​​the corresponding sub-region.

Images