An electronic paper master AOI defect detection system based on data analysis

Abstract

The present invention belongs to the technical field of electronic paper mother sheet detection, and specifically is an electronic paper mother sheet AOI defect detection system based on data analysis, comprising a server, an optical imaging module, an image processing module, a surface contamination impact assessment module and a defect precision identification module; the present invention performs high-definition imaging on the surface of the electronic paper mother sheet, and the surface contamination impact assessment module judges the surface contamination status of the electronic paper mother sheet based on the defect to be detected image and the surface standard image, and when generating a surface contamination low-impact signal, the defect on the surface of the electronic paper mother sheet is detected by the defect precision identification module and its defect performance status is reasonably evaluated, thereby ensuring the quality of the produced electronic paper mother sheet, and being able to accurately judge the detection performance of the electronic paper mother sheet AOI defect detection machine and perform progressive cause diagnosis when it is judged that its detection performance is poor, and having a high degree of intelligence and automation, which is conducive to ensuring the efficiency and accuracy of the detection and analysis process.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic paper master sheet detection, and in particular to an electronic paper master sheet AOI defect detection system based on data analysis. Background Art

[0002] The electronic paper mother sheet is one of the core components for manufacturing electronic paper and is the basic structure of electronic paper display devices. It is usually composed of a flexible or rigid substrate and a multi-layer structure such as a conductive layer, an insulating layer, and an electronic ink layer coated thereon. These layers are combined together through specific process steps to form a whole. During the production process of the electronic paper mother sheet, its surface defects need to be detected and identified to ensure the product quality of the electronic paper mother sheet produced;

[0003] At present, when performing defect detection on electronic paper mother sheets, it is difficult to reasonably evaluate the surface contamination status before defect detection to reduce the adverse effects on the accuracy of subsequent defect detection and identification results. It is also impossible to accurately judge the detection performance of the electronic paper mother sheet AOI defect detection machine and to perform progressive cause diagnosis when it is judged that its detection performance is poor. The intelligence and automation levels are low, which is not conducive to ensuring the efficiency and accuracy of the detection and analysis process.

[0004] In view of the above technical defects, a solution is now proposed. Summary of the invention

[0005] The purpose of the present invention is to provide an electronic paper master AOI defect detection system based on data analysis, which solves the problems that the prior art is difficult to reasonably evaluate the surface contamination status before defect detection, and is unable to accurately judge the detection performance of the electronic paper master AOI defect detection machine and perform progressive cause diagnosis when its detection performance is judged to be poor, and has a low degree of intelligence and automation.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] An electronic paper mother sheet AOI defect detection system based on data analysis comprises a server, an optical imaging module, an image processing module, a surface contamination impact assessment module and a defect precise identification module; the optical imaging module is located in the electronic paper mother sheet AOI defect detection machine, and uses a line scanning lens and a line scanning camera to perform high-definition imaging of the surface of the electronic paper mother sheet, collects the surface image of the electronic paper mother sheet and marks it as a defect to be inspected image, and sends the defect to be inspected image to the image processing module via the server; the image processing module performs denoising, enhancement and segmentation operations on the defect to be inspected image, and sends the processed defect to be inspected image of the electronic paper mother sheet to the surface contamination impact assessment module via the server;

[0008] The surface contamination impact assessment module retrieves the surface standard image of the electronic paper mother sheet from the server, judges the surface contamination status of the electronic paper mother sheet based on the defect inspection image of the electronic paper mother sheet and the surface standard image, generates a surface contamination high impact signal or a surface contamination low impact signal accordingly, and sends the surface contamination high impact signal to the supervision and early warning terminal via the server, and sends the surface contamination low impact signal to the defect accurate identification module via the server;

[0009] When the defect precise identification module receives a low-impact signal of surface contamination, it uses a deep learning algorithm to automatically analyze and identify the defect image to be inspected, detects the defects on the surface of the electronic paper mother sheet, and generates a defect quality inspection qualified signal or a defect quality inspection unqualified signal through a comprehensive defect evaluation analysis, and sends the defect quality inspection unqualified signal to the supervision and early warning end via the server; when the supervision and early warning end receives a high-impact signal of surface contamination or a defect quality inspection unqualified signal, it displays it and issues a corresponding early warning.

[0010] Furthermore, the specific operation process of the surface pollution impact assessment module includes:

[0011] Based on the defect inspection image of the electronic paper mother sheet and the surface standard image, the contaminated area on the surface of the electronic paper mother sheet is captured and marked as the surface contamination area; the area of ​​the corresponding surface contamination area is collected and marked as the surface contamination surface inspection value, the surface contamination surface inspection values ​​of all the surface contamination areas are summed up to obtain the surface contamination surface condition value, and the number of surface contamination areas is marked as the surface contamination count condition value; and the surface contamination surface condition value and the surface contamination count condition value are numerically compared with the preset surface contamination surface condition threshold and the preset surface contamination count condition threshold, respectively. If the surface contamination surface condition value or the surface contamination count condition value exceeds the corresponding preset threshold, a surface contamination high impact signal is generated.

[0012] Furthermore, if both the surface dirt surface condition value and the surface dirt number condition value do not exceed the corresponding preset threshold value, the deviation degree value between the actual color of the corresponding surface dirt area and the corresponding standard color is marked as the surface dirt color condition value, and the surface dirt surface inspection value of the corresponding surface dirt area and the surface dirt color condition value are numerically calculated to obtain the regional dirt condition value; the regional dirt condition value is numerically compared with the preset regional dirt condition threshold value, and if the regional dirt condition value exceeds the preset regional dirt condition threshold value, the corresponding surface dirt area is marked as a surface high dirt area; if there is a surface high dirt area on the electronic paper mother sheet, a surface dirt high impact signal is generated;

[0013] If there is no high surface pollution area on the electronic paper master sheet, the regional pollution values ​​of all surface pollution areas are summed and the average is taken to obtain the regional pollution value, and the regional pollution value, the surface pollution surface condition value and the surface pollution number condition value are numerically calculated to obtain the electronic paper master sheet surface inspection value; and the electronic paper master sheet surface inspection value is numerically compared with the preset electronic paper master sheet surface inspection threshold. If the electronic paper master sheet surface inspection value exceeds the preset electronic paper master sheet surface inspection threshold, a high impact signal of surface pollution is generated; if the electronic paper master sheet surface inspection value does not exceed the preset electronic paper master sheet surface inspection threshold, a low impact signal of surface pollution is generated.

[0014] Furthermore, the specific analysis process of the comprehensive defect evaluation analysis is as follows:

[0015] All defects are classified, and the preset thresholds of all inspection parameters of the corresponding types of defects are obtained. The actual detection values ​​of all inspection parameters of the corresponding defects are compared with the corresponding preset thresholds one by one. If there are inspection parameters exceeding the corresponding preset thresholds, the corresponding defects are marked as high-quality defects. If there are high-quality defects on the surface of the electronic paper mother sheet, a defect quality inspection failure signal is generated.

[0016] Furthermore, if there are no defects that have a high impact on the quality on the surface of the electronic paper master sheet, several detection blocks are segmented on the defect inspection image. If the detection block involves defects, the corresponding detection block is marked as a red-display block. The number of red-display blocks is obtained and the ratio is calculated with the total number of detection blocks to obtain a red-display detection value, and the red-display detection value is numerically compared with a preset red-display detection threshold. If the red-display detection value exceeds the preset red-display detection threshold, a defect quality inspection unqualified signal is generated; if the red-display detection value does not exceed the preset red-display detection threshold, a defect quality inspection qualified signal is generated.

[0017] Furthermore, the server is connected to the electronic paper master production tracing module in communication, and the electronic paper master production tracing module is used to set a monitoring period, collect the number of defect quality inspection unqualified signals generated during the monitoring period and mark it as a detection non-good value, and calculate the ratio of the detection non-good value to the number of electronic paper master sheets detected by the electronic paper master AOI defect detection machine during the monitoring period to obtain a detection non-analysis value;

[0018] The detection traceability value is obtained by numerically calculating the detection non-good value and the detection non-analysis value, and the detection traceability value is numerically compared with the preset detection traceability value range. If the detection traceability value is within the preset detection traceability value range, a traceability qualified signal is generated; if the detection traceability value is not within the preset detection traceability value range, a traceability unqualified signal is generated, and the traceability unqualified signal is sent to the supervision and early warning end via the server.

[0019] Furthermore, when the unqualified signal of production traceability is generated, the management personnel extract a number of electronic paper mother sheets from the products corresponding to the qualified signal of defective quality inspection produced during the monitoring period, and manually inspect the extracted electronic paper mother sheets to obtain the percentage value of defective products. Similarly, a number of electronic paper mother sheets are extracted from the products corresponding to the unqualified signal of defective quality inspection produced during the monitoring period, and the extracted electronic paper mother sheets are manually inspected to obtain the percentage value of qualified products.

[0020] The defective product proportion value and the qualified product proportion value are numerically compared with the preset defective product proportion threshold value and the preset qualified product proportion threshold value respectively. If the defective product proportion value or the qualified product proportion value exceeds the corresponding preset threshold value, it is judged that the defect detection effect of the electronic paper master sheet AOI defect inspection machine during the monitoring period is poor.

[0021] Furthermore, the server is in communication connection with the rest evaluation feedback module and the inspection area real-time monitoring and diagnosis module. When it is determined that the defect detection effect of the electronic paper master AOI defect detection machine during the monitoring period is poor, the rest evaluation feedback module analyzes the operation and rest status of the electronic paper master AOI defect detection machine, generates a rest abnormality signal or a rest normal signal through the analysis, and sends the rest abnormality signal or the rest normal signal to the supervision and early warning terminal via the server, and sends the rest normal signal to the inspection area real-time monitoring and diagnosis module via the server;

[0022] When the real-time monitoring and diagnosis module of the inspection area receives a normal rest signal, it obtains the area used for optical imaging in the electronic paper master AOI defect inspection machine and marks it as the target inspection area, analyzes the imaging environment performance of the target inspection area during the monitoring period, and generates an inspection area diagnosis failure signal or an identification diagnosis abnormality signal through the analysis, and sends the inspection area diagnosis failure signal or the identification diagnosis abnormality signal to the supervision and early warning terminal via the server.

[0023] Furthermore, the specific analysis process of the rest assessment feedback module is as follows:

[0024] The start time of the electronic paper master AOI defect inspection machine for the current operation is collected, the time difference between the current time and the start time of the current operation is calculated to obtain the actual inspection time, and the actual inspection time is numerically compared with the preset actual inspection time threshold. If the actual inspection time exceeds the preset actual inspection time threshold, a rest abnormal signal is generated;

[0025] If the actual inspection time does not exceed the preset actual inspection time threshold, the time difference between the start time of the current operation and the end time of the previous operation is calculated to obtain the rest time detection value, and the actual inspection time and the rest time detection value are numerically calculated to obtain the rest value; and the rest value is numerically compared with the preset rest threshold. If the rest value exceeds the preset rest threshold, a rest abnormal signal is generated; if the rest value does not exceed the preset rest threshold, a rest normal signal is generated.

[0026] Furthermore, the specific analysis process of the inspection area real-time monitoring and diagnosis module is as follows:

[0027] The vibration amplitude and vibration frequency of the part in contact with the electronic paper mother sheet in the target inspection area are collected in real time, and the vibration amplitude and vibration frequency are numerically compared with the preset vibration amplitude threshold and the preset vibration frequency threshold respectively. If the vibration amplitude or the vibration frequency exceeds the corresponding preset threshold, it is determined that the target inspection area is in a vibration warning state;

[0028] If the vibration amplitude and the vibration frequency do not exceed the corresponding preset thresholds, the temperature deviation data, humidity deviation data, light intensity deviation data, light color deviation data and dust detection data of the target inspection area are collected in real time, and the temperature deviation data, humidity deviation data, light intensity deviation data, light color deviation data and dust detection data are numerically calculated to obtain the target ring detection value; and the target ring detection value is numerically compared with the preset target ring detection threshold. If the target ring detection value exceeds the preset target ring detection threshold, it is determined that the target inspection area is in the ring detection warning state;

[0029] The total duration that the target inspection area is in the vibration warning state or the environmental inspection warning state during the monitoring period is collected and marked as the inspection area diagnosis time value, and the inspection area diagnosis time value is numerically compared with the preset inspection area diagnosis time threshold. If the inspection area diagnosis time value exceeds the preset inspection area diagnosis time threshold, an inspection area diagnosis failure signal is generated; if the inspection area diagnosis time value does not exceed the preset inspection area diagnosis time threshold, an identification diagnosis abnormality signal is generated.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] 1. In the present invention, the surface of the electronic paper mother sheet is imaged with high definition by the optical imaging module, and the surface contamination impact assessment module determines the surface contamination status of the electronic paper mother sheet based on the defect inspection image and the surface standard image of the electronic paper mother sheet, thereby ensuring the accuracy of the defect detection result. When generating the surface contamination low impact signal, the defect accurate identification module detects the defects on the surface of the electronic paper mother sheet, and generates a defect quality inspection qualified signal or a defect quality inspection unqualified signal through comprehensive defect evaluation and analysis, thereby ensuring the quality of the produced electronic paper mother sheet, with a high degree of intelligence and automation;

[0032] 2. In the present invention, the probability of abnormal detection of the electronic paper master AOI defect detection machine during the monitoring period is determined through the electronic paper master production traceability module, and a random inspection and analysis is performed to determine the defect detection effect of the electronic paper master AOI defect detection machine when an unqualified production traceability signal is generated. When it is determined that the defect detection effect of the electronic paper master AOI defect detection machine during the monitoring period is poor, the operation and repair status of the electronic paper master AOI defect detection machine is analyzed, and when a normal repair signal is generated, the imaging environment performance of the target inspection area during the monitoring period is analyzed. This can not only accurately determine the detection performance of the electronic paper master AOI defect detection machine, but also perform progressive cause diagnosis when it is determined that its detection performance is poor. It has a high degree of intelligence and automation, which is conducive to ensuring the efficiency and accuracy of the detection and analysis process. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;

[0034] Figure 1 This is a system block diagram of Embodiment 1 of the present invention;

[0035] Figure 2 This is a system block diagram of Embodiment 2, Embodiment 3 and Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] Embodiment 1: Figure 1 As shown, the electronic paper master sheet AOI (automatic optical inspection, i.e., visual inspection) defect detection system based on data analysis proposed by the present invention includes a server, an optical imaging module, an image processing module, a surface contamination impact assessment module, and a defect precision identification module, and the server is communicatively connected with the optical imaging module, the image processing module, the surface contamination impact assessment module, and the defect precision identification module;

[0038] The optical imaging module is located in the electronic paper mother sheet AOI defect inspection machine, and uses a line scan lens and a line scan camera to perform high-definition imaging of the surface of the electronic paper mother sheet, collects the surface image of the electronic paper mother sheet and marks it as a defective inspection image, and sends the defective inspection image to the image processing module via the server; the image processing module pre-processes the defective inspection image, including denoising, enhancement and segmentation operations, to improve the image quality, and sends the processed defective inspection image of the electronic paper mother sheet to the surface contamination impact assessment module via the server;

[0039] The surface pollution impact assessment module retrieves the surface standard image of the electronic paper mother sheet from the server, and judges the surface pollution status of the electronic paper mother sheet based on the defect image to be inspected and the surface standard image of the electronic paper mother sheet, and generates a surface pollution high impact signal or a surface pollution low impact signal accordingly, and sends the surface pollution high impact signal to the supervision and early warning end through the server. When the supervision and early warning end receives the surface pollution high impact signal, it displays it and issues a corresponding early warning. When the corresponding management personnel receive the early warning, they should promptly perform surface cleaning and other operations on the electronic paper mother sheet to reduce its adverse effects on the subsequent defect detection and identification process, and ensure the accuracy of the defect detection results; the specific operation process of the surface pollution impact assessment module is as follows:

[0040] Based on the defect inspection image of the electronic paper mother sheet and the surface standard image, the contaminated area on the surface of the electronic paper mother sheet is captured and marked as the surface contamination area; the area of ​​the corresponding surface contamination area is collected and marked as the surface contamination surface inspection value, the surface contamination surface inspection values ​​of all the surface contamination areas are summed up to obtain the surface contamination surface condition value, and the number of surface contamination areas is marked as the surface contamination number condition value; and the surface contamination surface condition value and the surface contamination number condition value are numerically compared with the preset surface contamination surface condition threshold value and the preset surface contamination number condition threshold value respectively. If the surface contamination surface condition value or the surface contamination number condition value exceeds the corresponding preset threshold value, it indicates that the surface contamination condition of the electronic paper mother sheet is serious, and a surface contamination high impact signal is generated;

[0041] If both the surface pollution surface condition value and the surface pollution number condition value do not exceed the corresponding preset threshold value, the deviation degree value between the actual color of the corresponding surface pollution area and the corresponding standard color is marked as the surface pollution color condition value. The larger the value of the surface pollution color condition value, the more serious the pollution degree of the corresponding surface pollution area is; the surface pollution surface inspection value WK and the surface pollution color condition value WS of the corresponding surface pollution area are numerically calculated by the formula WP=a1*WK+a2*WS to obtain the regional pollution condition value WP; wherein a1 and a2 are preset proportional coefficients, and the values ​​of a1 and a2 are both positive numbers; and the larger the value of the regional pollution condition value WP, the more serious the pollution condition of the corresponding surface pollution area is;

[0042] The regional pollution value WP is numerically compared with the preset regional pollution threshold. If the regional pollution value WP exceeds the preset regional pollution threshold, it indicates that the pollution condition of the corresponding surface pollution area is serious, and the corresponding surface pollution area is marked as a surface high pollution area; if there is a surface high pollution area on the electronic paper mother sheet, it indicates that the pollution condition of the corresponding electronic paper mother sheet is not good, and a surface pollution high impact signal is generated;

[0043] If there is no high-contamination area on the electronic paper mother sheet, the regional contamination values ​​of all surface contamination areas are summed and averaged to obtain the regional contamination value, and the regional contamination value BP, the surface contamination surface value BY and the surface contamination number value BF are numerically calculated by the formula BX=ew1*BP+ew2*BY+ew3*BF to obtain the electronic paper mother sheet surface inspection value BX; wherein ew1, ew2, ew3 are preset proportional coefficients, and the values ​​of ew1, ew2, ew3 are all greater than zero; and the larger the value of the electronic paper mother sheet surface inspection value BX, the more serious the surface contamination condition of the corresponding electronic paper mother sheet is as a whole;

[0044] The electronic paper master sheet surface inspection value BX is numerically compared with the preset electronic paper master sheet surface inspection threshold. If the electronic paper master sheet surface inspection value BX exceeds the preset electronic paper master sheet surface inspection threshold, it indicates that the surface contamination condition of the corresponding electronic paper master sheet is generally serious, and a surface contamination high impact signal is generated; if the electronic paper master sheet surface inspection value BX does not exceed the preset electronic paper master sheet surface inspection threshold, it indicates that the surface contamination condition of the corresponding electronic paper master sheet is generally good, and a surface contamination low impact signal is generated.

[0045] The surface contamination impact assessment module sends the surface contamination low impact signal to the defect precise identification module via the server. When the defect precise identification module receives the surface contamination low impact signal, it uses a deep learning algorithm to automatically analyze and identify the defect image to be inspected, detects the defects on the surface of the electronic paper mother sheet, and generates a defect quality inspection qualified signal or a defect quality inspection unqualified signal through comprehensive defect evaluation analysis, and sends the defect quality inspection unqualified signal to the supervision and early warning end via the server. When the supervision and early warning end receives the defect quality inspection unqualified signal, it displays it and issues a corresponding early warning, and the electronic paper mother sheet AOI defect detection machine conveys the electronic paper mother sheet corresponding to the defect quality inspection unqualified signal and the electronic paper mother sheet corresponding to the defect quality inspection qualified signal through different conveying channels for subsequent manual sampling, thereby reducing the difficulty of subsequent operations and ensuring the quality of the produced electronic paper mother sheet; the specific analysis process of the defect comprehensive evaluation analysis is as follows:

[0046] Classify all defects (including scratches, spots, bubbles, etc.), obtain preset thresholds of all inspection parameters of corresponding types of defects (for example, for scratches, the preset thresholds of the corresponding inspection parameters include scratch length threshold, scratch width threshold and scratch depth threshold, etc.), compare the actual detection values ​​of all inspection parameters of the corresponding defects with the corresponding preset thresholds one by one, if there are inspection parameters exceeding the corresponding preset thresholds, the corresponding defects are marked as high-quality defects, if there are high-quality defects on the surface of the electronic paper mother sheet, it indicates that the quality of the corresponding electronic paper mother sheet is poor, and a defect quality inspection failure signal is generated;

[0047] If there are no defects that have a high impact on the quality of the electronic paper mother sheet on the surface, several detection blocks are segmented on the defect inspection image. If the detection block involves defects, the corresponding detection block is marked as a red-display block; the number of red-display blocks is obtained and the ratio is calculated with the total number of detection blocks to obtain a red-display detection value, wherein the larger the value of the red-display detection value, the more serious the defect performance of the corresponding electronic paper mother sheet; and the red-display detection value is numerically compared with the preset red-display detection threshold value. If the red-display detection value exceeds the preset red-display detection threshold value, it indicates that the defect performance of the corresponding electronic paper mother sheet is more serious, and a defect quality inspection unqualified signal is generated; if the red-display detection value does not exceed the preset red-display detection threshold value, it indicates that the defect performance of the corresponding electronic paper mother sheet is not serious, and a defect quality inspection qualified signal is generated.

[0048] Embodiment 2: Figure 2 As shown, the difference between this embodiment and embodiment 1 is that the server is connected to the electronic paper master production tracing module in communication, and the electronic paper master production tracing module is used to set a monitoring period, preferably, the monitoring period is one hour; the number of defect quality inspection unqualified signals generated in the monitoring period is collected and marked as a detection non-good value, and the detection non-good value is calculated by ratio with the number of electronic paper master sheets detected by the electronic paper master AOI defect detection machine in the monitoring period to obtain a detection non-analysis value; the detection non-good value FY and the detection non-analysis value FK are numerically calculated by the formula FX=b1*FY / b2+b2*FK to obtain a detection production tracing value FX, wherein b1 and b2 are preset proportional coefficients, b2>b1>0; and the value of the detection production tracing value FX is too large or too small, which indicates that the probability of detection abnormality in the monitoring period is relatively high;

[0049] The detection traceability value FX is numerically compared with the preset detection traceability value range. If the detection traceability value FX is within the preset detection traceability value range, it indicates that the probability of detection abnormality during the monitoring period is small, and a traceability qualified signal is generated; if the detection traceability value FX is not within the preset detection traceability value range, it indicates that the probability of detection abnormality during the monitoring period is large, and a traceability unqualified signal is generated, and the traceability unqualified signal is sent to the supervision and early warning end via the server to remind the management personnel to promptly conduct random inspections of the electronic paper mother sheets that have been tested during the monitoring period.

[0050] Further, when generating a production traceability unqualified signal, the management personnel extract a number of electronic paper mother sheets from the products corresponding to the defect quality inspection qualified signals produced during the monitoring period, preferably no less than twenty; manually inspect the extracted electronic paper mother sheets to obtain a defective product ratio value (i.e., the product quantity ratio value of the corresponding electronic paper mother sheets that do not meet the requirements after manual inspection); similarly, extract a number of electronic paper mother sheets from the products corresponding to the defect quality inspection unqualified signals produced during the monitoring period, and manually inspect the extracted electronic paper mother sheets to obtain a qualified product ratio value (i.e., the product quantity ratio value of the corresponding electronic paper mother sheets that meet the requirements after manual inspection);

[0051] And the defective product ratio value and the qualified product ratio value are numerically compared with the preset defective product ratio threshold value and the preset qualified product ratio threshold value respectively. If the defective product ratio value or the qualified product ratio value exceeds the corresponding preset threshold value, it indicates that the error rate of the detection result of the electronic paper master sheet AOI defect detection machine during the monitoring period is relatively large, and then it is judged that the defect detection effect of the electronic paper master sheet AOI defect detection machine during the monitoring period is poor. When it is judged that its defect detection effect is poor, the management personnel are reminded to conduct cause investigation and tracing in time, and make matching improvement measures, so as to ensure the efficiency and accuracy of subsequent defect detection and identification.

[0052] Embodiment 3: Figure 2 As shown, the difference between this embodiment and embodiment 1 and embodiment 2 is that the server is connected to the rest evaluation feedback module in communication. When it is determined that the defect detection effect of the electronic paper master AOI defect detection machine is poor during the monitoring period, the rest evaluation feedback module analyzes the operation and rest status of the electronic paper master AOI defect detection machine, and generates a rest abnormality signal or a rest normal signal through analysis, so as to accurately judge the correlation between the poor operation accuracy of the electronic paper master AOI defect detection machine and its rest status, and send the rest abnormality signal or the rest normal signal to the supervision and early warning end through the server, and the supervision and early warning end displays the rest abnormality signal and issues a corresponding early warning. When the management personnel receive the corresponding early warning, they should suspend the work and rest the electronic paper master AOI defect detection machine as needed to ensure its performance and improve the accuracy of its defect detection results and the detection efficiency. The specific analysis process of the rest evaluation feedback module is as follows:

[0053] The start time of the current operation of the electronic paper master AOI defect detection machine is collected, and the time difference between the current time and the start time of the current operation is calculated to obtain the actual inspection time. The larger the value of the actual inspection time, the longer the duration of the current operation of the electronic paper master AOI defect detection machine is, and the less conducive it is to ensure the accuracy and efficiency of its defect detection results. The actual inspection time is compared with the preset actual inspection time threshold. If the actual inspection time exceeds the preset actual inspection time threshold, it indicates that the electronic paper master AOI defect detection machine is not rested in time, resulting in a high possibility that the accuracy of its image detection and recognition analysis results is poor, and a rest abnormal signal is generated;

[0054] If the actual inspection time does not exceed the preset actual inspection time threshold, the time difference between the start time of the current operation and the end time of the previous adjacent operation is calculated to obtain the rest time detection value, and the actual inspection time XS and the rest time detection value XP are numerically calculated by the formula XK=hy1*XS+hy2 / (XP+hy1) to obtain the rest value XK; wherein, hy1 and hy2 are preset proportional coefficients, hy2>hy1>0; and the rest value XK is numerically compared with the preset rest threshold, if the rest value XK exceeds the preset rest threshold, indicating that the possibility of poor accuracy of the image detection, recognition and analysis results of the electronic paper master AOI defect detection machine due to failure to rest in time is high, and a rest abnormal signal is generated; if the rest value XK does not exceed the preset rest threshold, indicating that the possibility of poor accuracy of the image detection, recognition and analysis results of the electronic paper master AOI defect detection machine due to failure to rest in time is low, and a rest normal signal is generated.

[0055] Embodiment 4: Figure 2As shown, the difference between this embodiment and embodiment 1, embodiment 2, and embodiment 3 is that the server is connected to the inspection area real-time monitoring and diagnosis module in communication, and the rest evaluation feedback module sends the normal rest signal to the inspection area real-time monitoring and diagnosis module via the server; when the inspection area real-time monitoring and diagnosis module receives the normal rest signal, it obtains the area for optical imaging in the electronic paper mother AOI defect detection machine and marks it as the target inspection area, analyzes the imaging environment performance of the target inspection area during the monitoring period, and generates an inspection area diagnosis unqualified signal or an identification diagnosis abnormal signal through analysis;

[0056] The unqualified diagnosis signal of the inspection area or the abnormal identification diagnosis signal is sent to the supervision and early warning terminal through the server. When the unqualified diagnosis signal of the inspection area is generated, the imaging environment supervision and control of the target inspection area shall be strengthened in time. When the abnormal identification diagnosis signal is generated, the corresponding identification analysis module software shall be optimized in time to reduce the management difficulty of the management personnel and quickly make targeted improvement measures. The high degree of intelligence and automation further ensures the performance of the electronic paper mother AOI defect detection machine; specifically, the specific analysis process of the real-time monitoring and diagnosis module of the inspection area is as follows:

[0057] The vibration amplitude and vibration frequency of the part in contact with the electronic paper mother sheet in the target inspection area are collected in real time. It should be noted that the larger the values ​​of the vibration amplitude and vibration frequency, the worse the stability of the corresponding part, which is more unfavorable to ensure the image quality of surface imaging, and the less effective it is to ensure the accuracy of subsequent product defect analysis results; the vibration amplitude and vibration frequency are numerically compared with the preset vibration amplitude threshold and the preset vibration frequency threshold respectively. If the vibration amplitude or vibration frequency exceeds the corresponding preset threshold, it is judged that the target inspection area is in a vibration warning state;

[0058] If the vibration amplitude and the vibration frequency do not exceed the corresponding preset thresholds, the temperature deviation data, humidity deviation data, light intensity deviation data, light color deviation data and dust detection data of the target inspection area are collected in real time, wherein the temperature deviation data is a data quantity value indicating the deviation value of the actual temperature in the target inspection area compared with the preset standard temperature, the humidity deviation data is a data quantity value indicating the deviation value of the actual humidity in the target inspection area compared with the preset standard humidity, the light intensity deviation data is a data quantity value indicating the deviation value of the actual light intensity in the target inspection area compared with the preset standard light intensity, the light color deviation data is a data quantity value indicating the deviation degree of the actual light color in the target inspection area compared with the preset standard light color, and the dust detection data is a data quantity value indicating the dust concentration in the target inspection area;

[0059] The temperature deviation data RF, humidity deviation data RS, light intensity deviation data RY, light color deviation data RX and dust detection data RP are numerically calculated by the formula RK=(q1*RF+q2*RS+q3*RY+q4*RX+q5*RP) / 5 to obtain the target ring detection value RK; wherein, q1, q2, q3, q4, q5 are preset proportional coefficients, and the values ​​of q1, q2, q3, q4, q5 are all positive numbers; and the larger the value of the target ring detection value RK, the worse the imaging environment performance of the target inspection area at the corresponding moment; the target ring detection value RK is numerically compared with the preset target ring detection threshold value. If the target ring detection value RK exceeds the preset target ring detection threshold value, it indicates that the imaging environment performance of the target inspection area at the corresponding moment is poor, and the target inspection area is judged to be in the ring detection warning state;

[0060] The total duration that the target inspection area is in the vibration warning state or the environmental inspection warning state during the monitoring period is collected and marked as the inspection area diagnosis time value, wherein the larger the inspection area diagnosis time value, the worse the inspection area condition of the target inspection area during the monitoring period, and the less conducive it is to ensuring the surface imaging quality of the electronic paper master sheet, and the inspection area diagnosis time value is numerically compared with the preset inspection area diagnosis time threshold. If the inspection area diagnosis time value exceeds the preset inspection area diagnosis time threshold, it indicates that the possibility of poor accuracy of the image detection, recognition and analysis results of the electronic paper master sheet AOI defect detection machine due to the target inspection area condition is relatively high, and an inspection area diagnosis failure signal is generated; if the inspection area diagnosis time value does not exceed the preset inspection area diagnosis time threshold, it indicates that the possibility of poor accuracy of the image detection, recognition and analysis results of the electronic paper master sheet AOI defect detection machine due to the target inspection area condition is relatively low, and an identification diagnosis abnormality signal is generated.

[0061] The working principle of the present invention is as follows: when in use, the surface of the electronic paper mother sheet is imaged with high definition through the optical imaging module, the surface image of the electronic paper mother sheet is collected and marked as a defect inspection image, the image processing module pre-processes the defect inspection image to improve the image quality, the surface contamination impact assessment module judges the surface contamination status of the electronic paper mother sheet based on the defect inspection image and the surface standard image of the electronic paper mother sheet, and reminds the corresponding management personnel to perform surface cleaning and other operations on the electronic paper mother sheet when a surface contamination high impact signal is generated, so as to reduce the adverse effect on the subsequent defect detection and identification process, and ensure the accuracy of the defect detection result; and the surface contamination impact assessment module sends the surface contamination low impact signal to the defect accurate identification module, the defect accurate identification module uses a deep learning algorithm to automatically analyze and identify the defect inspection image, detect the defects on the surface of the electronic paper mother sheet, and generate a defect quality inspection qualified signal or a defect quality inspection unqualified signal through defect comprehensive evaluation and analysis, and display the defect quality inspection unqualified signal when it is generated and issue a corresponding warning, so as to ensure the quality of the produced electronic paper mother sheet, and have a high degree of intelligence and automation.

[0062] The above formulas are all dimensionless and numerical calculations. The formula is a formula obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formula are set by technicians in this field according to actual conditions. The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the contents of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that technicians in the relevant technical field can understand and use the present invention well. The present invention is only limited by the claims and their full scope and equivalents.

Claims

1. An electronic paper master sheet AOI defect detection system based on data analysis, It is characterized in that It includes a server, an optical imaging module, an image processing module, a surface contamination impact assessment module and a defect precision identification module; the optical imaging module is located in the electronic paper mother sheet AOI defect detection machine, and uses a line scan lens and a line scan camera to perform high-definition imaging of the surface of the electronic paper mother sheet, collects the surface image of the electronic paper mother sheet and marks it as a defective image to be inspected, and sends the defective image to be inspected to the image processing module via the server; The image processing module performs denoising, enhancement and segmentation operations on the defective image to be inspected, and sends the processed defective image of the electronic paper mother sheet to the surface contamination impact assessment module via the server; The surface contamination impact assessment module retrieves the surface standard image of the electronic paper mother sheet from the server, judges the surface contamination status of the electronic paper mother sheet based on the defect inspection image of the electronic paper mother sheet and the surface standard image, generates a surface contamination high impact signal or a surface contamination low impact signal accordingly, and sends the surface contamination high impact signal to the supervision and early warning terminal via the server, and sends the surface contamination low impact signal to the defect accurate identification module via the server; When the defect accurate identification module receives the low impact signal of surface contamination, it uses the deep learning algorithm to automatically analyze and identify the defect image to be inspected, detects the defects on the surface of the electronic paper mother sheet, and generates a defect quality inspection qualified signal or a defect quality inspection unqualified signal through comprehensive defect evaluation and analysis, and sends the defect quality inspection unqualified signal to the supervision and early warning end through the server; when the supervision and early warning end receives the high impact signal of surface contamination or the defect quality inspection unqualified signal, it displays it and issues a corresponding early warning; The specific operation process of the surface pollution impact assessment module includes: Based on the defect inspection image of the electronic paper mother sheet and the surface standard image, the contaminated area on the surface of the electronic paper mother sheet is captured and marked as the surface contamination area; the area of ​​the corresponding surface contamination area is collected and marked as the surface contamination surface inspection value, the surface contamination surface inspection values ​​of all the surface contamination areas are summed up to obtain the surface contamination surface condition value, and the number of surface contamination areas is marked as the surface contamination number condition value; if the surface contamination surface condition value or the surface contamination number condition value exceeds the corresponding preset threshold value, a surface contamination high impact signal is generated; If both the surface dirt surface condition value and the surface dirt number condition value do not exceed the corresponding preset threshold value, the deviation degree value between the actual color of the corresponding surface dirt area and the corresponding standard color is marked as the surface dirt color condition value, and the surface dirt surface inspection value WK of the corresponding surface dirt area and the surface dirt color condition value WS are numerically calculated by the formula WP=a1*WK+a2*WS to obtain the regional dirt condition value WP; wherein a1 and a2 are preset proportional coefficients, and the values ​​of a1 and a2 are both positive numbers; if the regional dirt condition value exceeds the preset regional dirt condition threshold value, the corresponding surface dirt area is marked as a surface high dirt area; if there is a surface high dirt area on the electronic paper mother sheet, a surface dirt high impact signal is generated; If there is no high-contamination area on the electronic paper master sheet, the regional contamination values ​​of all surface contamination areas are summed and averaged to obtain the regional contamination value, and the regional contamination value BP, the surface contamination surface value BY and the surface contamination number value BF are numerically calculated by the formula BX=ew1*BP+ew2*BY+ew3*BF to obtain the electronic paper master sheet surface inspection value BX; wherein ew1, ew2, and ew3 are preset proportional coefficients, and the values ​​of ew1, ew2, and ew3 are all greater than zero; if the electronic paper master sheet surface inspection value exceeds the preset electronic paper master sheet surface inspection threshold, a high-impact signal of surface contamination is generated; if the electronic paper master sheet surface inspection value does not exceed the preset electronic paper master sheet surface inspection threshold, a low-impact signal of surface contamination is generated.

2. The electronic paper master AOI defect detection system based on data analysis according to claim 1, It is characterized in that The specific analysis process of comprehensive defect assessment and analysis is as follows: All defects are classified, and the preset thresholds of all inspection parameters of the corresponding types of defects are obtained. The actual detection values ​​of all inspection parameters of the corresponding defects are compared with the corresponding preset thresholds one by one. If there are inspection parameters exceeding the corresponding preset thresholds, the corresponding defects are marked as high-quality defects. If there are high-quality defects on the surface of the electronic paper mother sheet, a defect quality inspection failure signal is generated.

3. The electronic paper master sheet AOI defect detection system based on data analysis according to claim 2, It is characterized in that If there are no defects that have a high impact on the quality on the surface of the electronic paper mother sheet, several detection blocks are segmented on the defect inspection image. If the detection block involves defects, the corresponding detection block is marked as a red-display block. The number of red-display blocks is obtained and the ratio is calculated with the total number of detection blocks to obtain a red-display detection value. If the red-display detection value exceeds the preset red-display detection threshold, a defect quality inspection unqualified signal is generated; if the red-display detection value does not exceed the preset red-display detection threshold, a defect quality inspection qualified signal is generated.

4. The electronic paper master AOI defect detection system based on data analysis according to claim 1, It is characterized in that The server is in communication connection with the electronic paper master production tracing module, which is used to set a monitoring period, collect the number of defect quality inspection unqualified signals generated during the monitoring period and mark it as a detection non-good value, and calculate the ratio of the detection non-good value to the number of electronic paper master sheets detected by the electronic paper master AOI defect detection machine during the monitoring period to obtain a detection non-analysis value; The detection value is obtained by numerically calculating the detection non-good value and the detection non-analysis value. If the detection value is within the preset detection value range, a qualified signal is generated. If the detected traceability value is not within the preset traceability value range, a traceability failure signal is generated, and the traceability failure signal is sent to the supervision and early warning end via the server.

5. The electronic paper master AOI defect detection system based on data analysis according to claim 4, It is characterized in that When generating a production traceability unqualified signal, the management personnel will extract a number of electronic paper master sheets from the products corresponding to the defect quality inspection qualified signals produced during the monitoring period, and manually inspect the extracted electronic paper master sheets to obtain the defective product ratio. Similarly, extract a number of electronic paper master sheets from the products corresponding to the defect quality inspection unqualified signals produced during the monitoring period, and manually inspect the extracted electronic paper master sheets to obtain the qualified product ratio. If the defective product ratio or the qualified product ratio exceeds the corresponding preset threshold, it is judged that the defect detection effect of the electronic paper master AOI defect detection machine during the monitoring period is poor.

6. The electronic paper master AOI defect detection system based on data analysis according to claim 4, It is characterized in that The server is in communication connection with the rest evaluation feedback module and the inspection area real-time monitoring and diagnosis module. When it is judged that the defect detection effect of the electronic paper master AOI defect detection machine during the monitoring period is poor, the rest evaluation feedback module analyzes the operation and rest status of the electronic paper master AOI defect detection machine, generates a rest abnormality signal or a rest normal signal through the analysis, and sends the rest abnormality signal or the rest normal signal to the supervision and early warning terminal via the server, and sends the rest normal signal to the inspection area real-time monitoring and diagnosis module via the server; When the real-time monitoring and diagnosis module of the inspection area receives a normal rest signal, it obtains the area used for optical imaging in the electronic paper master AOI defect inspection machine and marks it as the target inspection area, analyzes the imaging environment performance of the target inspection area during the monitoring period, and generates an inspection area diagnosis failure signal or an identification diagnosis abnormality signal through the analysis, and sends the inspection area diagnosis failure signal or the identification diagnosis abnormality signal to the supervision and early warning terminal via the server.

7. The electronic paper master AOI defect detection system based on data analysis according to claim 6, It is characterized in that The specific analysis process of the rest assessment feedback module is as follows: The start time of the electronic paper master AOI defect inspection machine is collected, and the time difference between the current time and the start time of the current operation is calculated to obtain the actual inspection time. If the actual inspection time exceeds the preset actual inspection time threshold, a rest abnormal signal is generated; If the actual inspection time does not exceed the preset actual inspection time threshold, the time difference between the start time of the current operation and the end time of the previous operation is calculated to obtain the rest time detection value, and the actual inspection time XS and the rest time detection value XP are numerically calculated by the formula XK=hy1*XS+hy2 / (XP+hy1) to obtain the rest value XK; wherein, hy1 and hy2 are preset proportional coefficients, hy2>hy1>0; if the rest value exceeds the preset rest threshold, a rest abnormality signal is generated; if the rest value does not exceed the preset rest threshold, a rest normal signal is generated.

8. The electronic paper master AOI defect detection system based on data analysis according to claim 6, It is characterized in that The specific analysis process of the inspection area real-time monitoring and diagnosis module is as follows: The vibration amplitude and vibration frequency of the part in contact with the electronic paper mother sheet in the target inspection area are collected in real time. If the vibration amplitude or vibration frequency exceeds the corresponding preset threshold, it is judged that the target inspection area is in a vibration warning state; if the vibration amplitude and vibration frequency do not exceed the corresponding preset threshold, the temperature deviation data, humidity deviation data, light intensity deviation data, light color deviation data and dust detection data of the target inspection area are collected in real time, and the temperature deviation data, humidity deviation data, light intensity deviation data, light color deviation data and dust detection data are numerically calculated to obtain the target inspection ring value; If the target ring detection value exceeds the preset target ring detection threshold, the target detection area is judged to be in the ring detection warning state; The total duration that the target inspection area is in the vibration warning state or the environmental inspection warning state during the monitoring period is collected and marked as the inspection area diagnosis time value. If the inspection area diagnosis time value exceeds the preset inspection area diagnosis time threshold, an inspection area diagnosis failure signal is generated; if the inspection area diagnosis time value does not exceed the preset inspection area diagnosis time threshold, an identification diagnosis abnormality signal is generated.

Images