A fast panel detection method based on camera front and back depth of field

Abstract

This invention discloses a rapid panel inspection method based on the camera's front and rear depth of field. This method replaces the previous method of capturing images while testing and then releasing the machine until the test is complete, with a method that releases the machine only after the camera has captured all images. This method utilizes the machine's mobile photo-taking time for inspection, significantly reducing the time the machine spends in the equipment. This method utilizes front and rear depth of field sliders for defect comparison, achieving higher accuracy than previous methods, making it less likely to miss small defects and less likely to falsely report large defects.

Description

Technical Field

[0001] The present invention relates to the technical field of camera panel detection, and in particular to a panel rapid detection method based on the front and rear depth of field of a camera. Background Art

[0002] The existing panel AOI inspection process and basic algorithm are as follows: a) After the product machine is delivered to the station by belt / roller, the photoelectric sensor switch of the line PLC is triggered, which then sends a start test signal to the test machine computer via the serial port; b) After receiving the test signal, the computer sends instructions to the image signal generator (K8266) to send white, black, red, green, and blue image signals in sequence; c) Two 65-megapixel main cameras and four 25-megapixel auxiliary cameras then capture each image; d) Detection Algorithm 1: Dust Removal Algorithm. The machine utilizes the characteristic of dust to reflect light under strong light. When the display is in sleep mode, two white light bars provide light to reveal the location of dust, ignoring the coordinates of the dust in subsequent image inspections. Detection Algorithm 2: White Screen Detection: This method uses a black background and the brightness of the display screen to locate and crop the product. The original image is cropped based on the product's coordinates and partitioned into a nine-square grid. The pixel mean of each partition is calculated. This mean is then applied to a preset threshold and binarized, eroded, or expanded to reveal any pixels with abrupt color changes. A preliminary defect determination is made by counting the number of abrupt pixels and their connected area. After a preliminary defect determination, the coordinates are recorded and returned to the original image. The pixel mean of the defective area is calculated and then compared with the absolute value of the pixel mean of the nine-square grid to determine whether the defect is present. f) Detection Algorithm 3: Black Screen Detection: Using the product coordinates located by white screen detection, the black screen image is cropped. The black and white colors are inverted using binary inversion. The subsequent detection algorithm is similar to the white screen algorithm. g) Detection Algorithm 4: Dark Screen Detection: The basic algorithm is the same as the black screen algorithm, but the debugging parameters are different. h) After the test is completed, the machine releases the product and sends it to the return flow judgment station based on the test results. Based on the test results, it is judged whether the product is qualified. If qualified, it will continue to flow to the next station. If unqualified, it will be marked with a return flow label and go offline.

[0003] It can be seen that the test time of the Panel AOI machine in the existing technology consists of: a) the time for the product to flow into the equipment, 3s; b) the time the product stays inside the equipment; c) the time it takes for the product to completely flow out of the equipment.

[0004] In the era of Industry 4.0, companies are introducing a large number of automated equipment for product quality inspection: panel inspection, HDMI and DP interface screen inspection, and EDID burn-in inspection. These equipment replaces manual inspections, accelerating unmanned operations in mid-line inspection processes, reducing labor costs, and alleviating labor shortages. Therefore, these automated equipment must meet the following technical requirements and overcome the following technical deficiencies: 1. Excessively long inspection time (CT): Current inspection machines have an inspection speed of 20s–24s (the number of product defects affects inspection time). The speeds required for different products vary, ranging from 16s–18s for 24-inch products, 18s–22s for 27-inch products, and 22s–26s for 32-inch products. Therefore, the current test CT and production fault tolerance CT of inspection machines cannot meet these requirements in high-speed, high-capacity production lines. 2. Detection accuracy is limited to cover shallow, small dark spots: Current machine inspections are mostly based on fixed parameters such as area and pixel value, making them neither universal nor flexible. This can easily lead to false positives and omissions, leading to a lack of trust in the machines by production and quality departments. 3. Product inspection specifications require multiple revisions for the same product: Because different raw material suppliers produce different quality requirements, and sometimes different work orders for the same product may have different suppliers, the current inspection specification settings sometimes require frequent online maintenance by engineers, which can lead to untimely maintenance. Summary of the Invention

[0005] The purpose of the present invention is to provide a panel rapid detection method based on the front and rear depth of field of the camera, improve and reduce the test time to increase output, optimize the accuracy of the test algorithm to reduce false alarms and missed detections of defects, transform the product defect characteristics from concrete to digital abstraction, and use its mathematical characteristics to set parameters to reduce the number of times the inspection specification is maintained.

[0006] The technical solution adopted in the present invention is:

[0007] A panel rapid detection method based on the front and rear depth of field of a camera comprises the following steps:

[0008] Step 1: The product to be tested flows into the test platform and is transferred to the test position;

[0009] Step 2: Determine whether it is the first test product; if so, proceed to step 4; otherwise, proceed to step 3;

[0010] Step 3: Determine whether the test end signal of the previous product to be tested is received; if so, execute step 4; otherwise, wait for the set time and execute step 3;

[0011] Step 4: Create a test thread for the product to be tested according to the order stored in the List container;

[0012] Step 5: The testing equipment on the testing platform takes a picture of the product to be tested and tests the product to be tested at the same time;

[0013] Step 6: After all tests are completed, the corresponding test results are generated and the products to be tested are transferred to the outflow station;

[0014] Step 7: Upload the test results, generate a test end signal, and execute step 1.

[0015] Furthermore, the detection in step 5 includes the following steps:

[0016] Step 5-1: During the test, the entire panel of the camera is divided into nine areas and the pixel mean of each area is obtained;

[0017] Step 5-2: Set a foreground rectangle slider and slide the foreground rectangle slider from left to right and from top to bottom in the respective area according to the first sliding step length to traverse the entire area. Each time the foreground rectangle slider is slid, the average value of the pixels in the foreground rectangle slider is obtained.

[0018] Furthermore, the area of ​​the foreground rectangular slider is 1 / 20 of the area of ​​the corresponding region.

[0019] Furthermore, the first sliding step is set to 3 pixel values.

[0020] Step 5-3: Set the background rectangle slider to detect the background rectangle slider and slide it from left to right and from top to bottom in the foreground rectangle slider according to the second sliding step to traverse the entire foreground rectangle slider. Each time it slides, the average pixel value in the background rectangle slider is obtained.

[0021] Furthermore, the area of ​​the background rectangular slider is 1 / 3 of the foreground rectangular slider.

[0022] Furthermore, the second sliding step is set to 1 pixel.

[0023] In step 5-4, the absolute value of the difference between the pixel mean values ​​of the foreground rectangular slider and the background rectangular slider is determined to preliminarily determine whether the position of the background rectangular slider is a defect position. If so, the foreground rectangular slider is binarized based on the pixel mean value of the background rectangular slider to calculate the defect point area, and the defect area is determined as required.

[0024] Furthermore, the specific steps for determining whether the background rectangle slider position is a defect position in step 5-4 are as follows:

[0025] Step 5-4-11. The area of ​​the foreground rectangle slider is larger than that of the background rectangle slider (about three times the area of ​​the background rectangle slider, for example, the foreground rectangle slider is 100*100 and the background rectangle slider is 30*30. It can be set according to actual conditions. The smaller the area, the higher the test accuracy and the longer the test time).

[0026] Step 5-4-12, calculate the pixel average of all pixels in the foreground rectangular slider area and the pixel average of the background rectangular slider (pixel value range is 0-255);

[0027] Step 5-4-13: Take the absolute value of the difference between the pixel means of the foreground rectangle slider and the background rectangle slider. When the absolute value exceeds the preset parameter value, the upper left corner coordinate of the background rectangle slider is used as the possible defect coordinate. Subsequently, the background rectangle slider is further analyzed for the defect area and pixel value size of the defect.

[0028] Furthermore, the specific steps for determining the defect area according to the requirements in step 5-4 are as follows:

[0029] Step 5-4-21. Calculate the area of ​​defects in the background rectangle slider: Use the connected domain algorithm (to determine whether dark-colored pixels are connected, and calculate the area of ​​all pixels whose values ​​are less than the average value of the background rectangle slider).

[0030] Step 5-4-22, dark spot and bright spot defect judgment: The defect position whose pixel value is smaller than the preset parameter value of the device (the parameter set when the template configuration is established) and whose area is larger than the preset point defect area parameter and smaller than the preset spot defect area parameter is judged as a dark spot or bright spot defect;

[0031] Step 5-4-23, dark spot and light spot defect judgment: The pixel value is less than the preset parameter value of the device (the parameter set when the template configuration is established), and the area is greater than the preset spot defect area parameter. Among them, the parameters in the template are based on the customer's requirements. The defect position is actually measured on the screen using a dot-line gauge to determine whether it is a dark spot or light spot defect;

[0032] Furthermore, after step 7 is completed, the machine releases the product to be tested and sends the test results to the reflow judgment station; the product is judged whether it is qualified based on the test results. If it is qualified, it continues to flow to the next station. If it is unqualified, it is marked with a reflow label and goes offline.

[0033] Furthermore, in step 7, when the program crashes due to itself or external factors, resulting in a lack of test results for the product under test, the corresponding product under test is determined to be an NG product to avoid quality loss, and a test end signal is generated.

[0034] This invention utilizes the above technical solution, replacing the previous method of capturing images while testing and then releasing the machine until the test is complete. This method utilizes the machine's mobile photo-taking time for testing, significantly reducing the time the machine spends in the equipment. This method uses front and rear depth of field sliders for defect comparison, achieving higher accuracy than previous methods, making it less likely to miss small defects and less likely to falsely report large defects. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments;

[0036] Figure 1 It is a flow chart of a panel detection method in the prior art;

[0037] Figure 2 This is a flow chart of a method for rapid panel detection based on camera front and rear depth of field according to the present invention;

[0038] Figure 3 Schematic diagram of the nine-square grid area division of the panel in the present invention;

[0039] Figure 4 This is a schematic diagram of the foreground rectangular slider of the present invention;

[0040] Figure 5 This is a schematic diagram of setting the background rectangular slider of the present invention. Implementation Method

[0041] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0042] like Figures 2 to 5 As shown, the present invention discloses a panel rapid detection method based on the front and rear depth of field of a camera, which includes the following steps:

[0043] Step 1: The product to be tested flows into the test platform and is transferred to the test position;

[0044] Step 2: Determine whether it is the first test product; if so, proceed to step 4; otherwise, proceed to step 3;

[0045] Step 3: Determine whether the test end signal of the previous product to be tested is received; if so, execute step 4; otherwise, wait for the set time and execute step 3;

[0046] Step 4: Create a test thread for the product to be tested according to the order stored in the List container;

[0047] Step 5: The testing equipment on the testing platform takes a picture of the product to be tested and tests the product to be tested at the same time;

[0048] Step 6: After all tests are completed, the corresponding test results are generated and the products to be tested are transferred to the outflow station;

[0049] Step 7: Upload the test results, generate a test end signal, and execute step 1.

[0050] Furthermore, the detection in step 5 includes the following steps:

[0051] Step 5-1, during the test, the entire panel of the camera is divided into nine regions 1 and the pixel mean of each region 1 is obtained;

[0052] Specifically, the nine-grid area division: LCD displays rely on a backlight panel to illuminate the display. Due to factors such as the outer casing and the arrangement of liquid crystals on the LCD panel, the brightness of the entire panel is not uniform, with dark edges and bright center. Based on this, the entire panel is divided into nine areas 1 during testing and the pixel average of each area 1 is obtained.

[0053] Step 5-2: Set a foreground rectangular slider 2 and slide the foreground rectangular slider 2 in the respective area 1 from left to right and from top to bottom according to a first sliding step length to traverse the entire area 1. Each time the foreground rectangular slider 2 slides, the average value of the pixels in the foreground rectangular slider 2 is obtained.

[0054] Furthermore, the area of ​​the foreground rectangular slider 2 is 1 / 20 of the area of ​​the corresponding region.

[0055] Furthermore, the first sliding step is set to 3 pixel values.

[0056] Specifically, if Figure 1 As shown: b) Foreground setting: Set the foreground rectangular slider 2, whose area is usually 1 / 20 of the region area, and slide it from left to right and from top to bottom in each area to traverse the entire area. The sliding step is usually set to 3 pixels, and the average pixel value in the foreground rectangular slider 2 is obtained once each sliding.

[0057] Step 5-3: Set the background rectangular slider 3 to be detected and slide the background rectangular slider 3 from left to right and from top to bottom in the foreground rectangular slider according to the second sliding step to traverse the entire foreground rectangular slider. Each time the slider slides, the pixel average value in the background rectangular slider 3 is obtained.

[0058] Furthermore, the area of ​​the background rectangular slider 3 is 1 / 3 of the foreground rectangular slider 2 .

[0059] Furthermore, the second sliding step is set to 1 pixel.

[0060] like Figure 2As shown: c) Background setting: Set the background detection rectangular slider 3, whose area is usually 1 / 3 of the foreground, so that it slides from left to right and from top to bottom in the foreground rectangular slider to traverse the entire foreground. The sliding step size is usually set to 1, and the average pixel value in the background rectangular slider 3 is obtained every time it slides.

[0061] In step 5-4, the absolute value of the difference between the pixel mean values ​​of the foreground rectangular slider and the background rectangular slider is determined to preliminarily determine whether the position of the background rectangular slider is a defect position. If so, the foreground rectangular slider is binarized based on the pixel mean value of the background rectangular slider to calculate the defect point area, and the defect area is determined as required.

[0062] like Figure 3 As shown: d) Defect judgment setting: By judging the absolute value of the difference between the pixel mean of the foreground rectangle slider and the background rectangle slider, it is preliminarily judged whether the position of the background rectangle slider is a defect position. If it is, the foreground rectangle slider is binarized according to the pixel mean of the background rectangle slider to calculate the area of ​​the defect point, and the defect area is judged according to the requirements.

[0063] Furthermore, the specific steps for determining whether the background rectangle slider position is a defect position in step 5-4 are as follows:

[0064] Step 5-4-11. The area of ​​the foreground rectangle slider is larger than that of the background rectangle slider (about three times the area of ​​the background rectangle slider, for example, the foreground rectangle slider is 100*100 and the background rectangle slider is 30*30. It can be set according to actual conditions. The smaller the area, the higher the test accuracy and the longer the test time).

[0065] Step 5-4-12, calculate the pixel average of all pixels in the foreground rectangular slider area and the pixel average of the background rectangular slider (pixel value range is 0-255);

[0066] Step 5-4-13: Take the absolute value of the difference between the pixel means of the foreground rectangle slider and the background rectangle slider. When the absolute value exceeds the preset parameter value, the upper left corner coordinate of the background rectangle slider is used as the possible defect coordinate. Subsequently, the background rectangle slider is further analyzed for the defect area and pixel value size of the defect.

[0067] Furthermore, the specific steps for determining the defect area according to the requirements in step 5-4 are as follows:

[0068] Step 5-4-21. Calculate the area of ​​defects in the background rectangle slider: Use the connected domain algorithm (to determine whether dark-colored pixels are connected, and calculate the area of ​​all pixels whose values ​​are less than the average value of the background rectangle slider).

[0069] Step 5-4-22, dark spot and bright spot defect judgment: The defect position whose pixel value is smaller than the preset parameter value of the device (the parameter set when the template configuration is established) and whose area is larger than the preset point defect area parameter and smaller than the preset spot defect area parameter is judged as a dark spot or bright spot defect;

[0070] Step 5-4-23, Dark Spot and Light Spot Defect Judgment: The pixel value is less than the device's preset parameter value (the parameter set when creating the template configuration), and the area is greater than the preset spot defect area parameter. The parameters in the template are based on the customer's requirements, and the defect location is determined by actual measurement of the screen using a dot-line gauge. The parameters in the template are based on the customer's requirements, and the defect location is determined to be a dark spot or light spot defect.

[0071] Furthermore, after step 7 is completed, the machine releases the product to be tested and sends the test results to the reflow judgment station; the product is judged whether it is qualified based on the test results. If it is qualified, it continues to flow to the next station. If it is unqualified, it is marked with a reflow label and goes offline.

[0072] Furthermore, in step 7, when the program crashes due to itself or external factors, resulting in a lack of test results for the product under test, the corresponding product under test is determined to be an NG product to avoid quality loss, and a test end signal is generated.

[0073] The specific principles of the present invention are described in detail below. Compared with the prior art, the present invention has the following technical improvements:

[0074] 1. Optimization of detection time CT: The test composition of the machine consists of three parts: product inflow equipment, equipment testing, and product outflow equipment. By optimizing "equipment testing" and "product outflow equipment", the optimization of shortening the time CT can be achieved.

[0075] a) The original program test logic is as follows Figure 1 As shown;

[0076] b) The test logic of the present invention uses multi-threading technology to overlap the equipment detection time with the product flow and photography time to achieve the effect of shortening the time CT. Figure 2 As shown, the special case processing logic is as follows:

[0077] i. Scenario 1: The production speed has increased, making the machine flow and photography time much shorter than the testing time. As a result, the next product has been photographed and the testing thread demand has been initiated before the previous product has been tested. If this continues, the computer memory will be filled up and the program will crash.

[0078] Solution to Case 1: Avoid opening a thread for each product. Use a List container to store the test start signal of each product. Only after receiving the corresponding test end signal will the test thread of the next machine be opened according to the order stored in the List container.

[0079] ii. Case 2: The program crashes due to internal or external factors, resulting in a lack of test data for a particular product.

[0080] Solution for Case 2: If a product is shut down before the test is complete, the test data will not be sent to the return flow judgment station. In this case, the return flow test station will directly judge the product as an NG product to avoid quality loss.

[0081] 2. Optimizing detection accuracy: a) Nine-grid area division: LCD displays rely on backlight panels to illuminate the display. Due to factors such as the housing and the arrangement of liquid crystals on the LCD panel, the brightness of the entire panel is uneven, with dark edges and bright center. Based on this, the entire panel is divided into nine areas during testing and the pixel average value of each area is obtained.

[0082] like Figure 1 As shown: b) Foreground setting: Set the foreground rectangular slider 2, whose area is usually 1 / 20 of the region area, and slide it from left to right and from top to bottom in each area to traverse the entire area. The sliding step is usually set to 3 pixels, and the average pixel value in the foreground rectangular slider 2 is obtained once each sliding.

[0083] like Figure 2 As shown: c) Background setting: Set the background detection rectangular slider 3, whose area is usually 1 / 3 of the foreground, so that it slides from left to right and from top to bottom in the foreground rectangular slider to traverse the entire foreground. The sliding step size is usually set to 1, and the average pixel value in the background rectangular slider 3 is obtained every time it slides.

[0084] like Figure 3 As shown: d) Defect judgment setting: By judging the absolute value of the difference in the pixel mean between the foreground rectangular slider and the background rectangular slider, it is preliminarily judged whether the position of the background rectangular slider is a defect position. If so, the foreground rectangular slider is binarized according to the pixel mean of the background rectangular slider to calculate the area of ​​the defect point, and the defect area is judged according to the requirements.

[0085] Specifically, Figure 3 For a whole screen, divide the whole screen into 9 areas (because the backlight panel of the backlit LED display cannot evenly cover the entire screen when illuminating, the average brightness of each area is different. If the foreground and background sliders are slid directly across the entire screen, it will cause extremely high errors). Figure 4The 1 represents an area in the entire screen, and 2 represents the foreground slider. The foreground slider will slide left and down in the direction of the arrow until the entire area is traversed. Figure 5 1 represents an area on the entire screen, 2 represents the foreground slider, and 3 represents the background slider. The background slider will also traverse the entire foreground slider in the direction of the arrow. Once the background slider has traversed the foreground slider, the foreground slider will slide once in the same direction, and then repeat the same process until the entire area has been traversed. The slider sliding rule is: first slide left to the end, then return to the beginning of the next row, and then continue sliding left until the last row.

[0086] This invention utilizes the above technical solution, replacing the previous method of capturing images while testing and then releasing the machine until the test is complete. This method utilizes the machine's mobile photo-taking time for testing, significantly reducing the time the machine spends in the equipment. This method uses front and rear depth of field sliders for defect comparison, achieving higher accuracy than previous methods, making it less likely to miss small defects and less likely to falsely report large defects.

[0087] Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. In the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the detailed description of the embodiments of the present application is not intended to limit the scope of the application for protection, but merely represents the selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

Claims

1. A panel rapid detection method based on the front and rear depth of field of a camera, characterized by: It includes the following steps: Step 1: The product to be tested flows into the test platform and is transferred to the test position; Step 2: Determine whether it is the first test product; If yes, go to step 4; Otherwise, go to step 3; Step 3: Determine whether the test end signal of the previous product to be tested is received; if so, proceed to step 4; Otherwise, wait for the set time and go to step 3; Step 4: Create a test thread for the product to be tested according to the order stored in the List container; Step 5: The testing equipment on the testing platform takes a picture of the product to be tested and tests the product to be tested at the same time; The test involves the following steps: Step 5-1: During the test, the entire panel of the camera is divided into nine areas and the pixel mean of each area is obtained; Step 5-2: Set a foreground rectangle slider and slide the foreground rectangle slider from left to right and from top to bottom in the respective area according to the first sliding step length to traverse the entire area. Each time the foreground rectangle slider is slid, the average value of the pixels in the foreground rectangle slider is obtained. Step 5-3: Setting a detection background rectangle slider and sliding the background rectangle slider from left to right and from top to bottom within the foreground rectangle slider according to a second sliding step length to traverse the entire foreground rectangle slider, obtaining a pixel average value within the background rectangle slider with each sliding; Step 5-4: Preliminarily determine whether the position of the background rectangular slider is a defect by determining the absolute value of the difference in pixel mean between the foreground rectangular slider and the background rectangular slider. If so, binarize the foreground rectangular slider based on the pixel mean of the background rectangular slider to calculate the defect area, and determine the defect area based on the requirements. The specific steps for determining whether the background rectangle slider position is a defect position in step 5-4 are as follows: Step 5-4-11, set the area of ​​the foreground rectangle slider to be no less than 3 times the area of ​​the background rectangle slider; Step 5-4-12, calculate the pixel average of all pixels in the area of ​​the foreground rectangular slider and the pixel average of the background rectangular slider; Step 5-4-13, take the absolute value of the difference between the pixel mean values ​​of the foreground rectangle slider and the background rectangle slider. If the absolute value exceeds the preset parameter value, the upper left corner coordinate of the background rectangle slider is used as the possible defect coordinate. The background rectangle slider is then further analyzed for defect area and pixel value size. The specific steps for determining the defect area based on requirements in step 5-4 are as follows: Step 5-4-21. Calculate the area of ​​defects in the background rectangle slider: Use the connected domain algorithm to determine whether dark-colored pixels are connected, and calculate the area of ​​all pixel regions whose pixel values ​​are less than the average value of the background rectangle slider pixels. Step 5-4-22, dark spot or bright spot defect judgment: The defect position whose pixel value is smaller than the preset parameter value of the device and whose area is larger than the preset point defect area parameter and smaller than the preset spot defect area parameter is judged as a dark spot or bright spot defect; Step 5-4-23, dark spot or light spot defect judgment: The defect position with a pixel value smaller than the preset parameter value of the device and an area larger than the preset spot defect area parameter is judged as a dark spot or light spot defect; Step 6: After all tests are completed, the corresponding test results are generated and the products to be tested are transferred to the outflow station; Step 7: Upload the test results, generate a test end signal, and execute step 1.

2. The method for rapid panel inspection based on camera front and back depth of field according to claim 1, characterized in that: The area of ​​the foreground rectangular slider is 1 / 20 of the area of ​​the corresponding region.

3. The method for rapid panel inspection based on camera front and back depth of field according to claim 1, characterized in that: The first sliding step is set to 3 pixels.

4. The method for rapid panel inspection based on camera front and back depth of field according to claim 1, characterized in that: The area of ​​the background rectangle slider is 1 / 3 of the foreground rectangle slider.

5. The method for rapid panel inspection based on camera front and back depth of field according to claim 1, characterized in that: The second sliding step is set to 1 pixel.

6. The method for rapid panel inspection based on camera front and back depth of field according to claim 1, characterized in that: After step 7, the machine releases the product to be tested and sends the test results to the return flow judgment station; the product is judged whether it is qualified based on the test results. If it is qualified, it continues to flow to the next station. If it is unqualified, it is marked with a return flow label and goes offline.

7. The method for rapid panel inspection based on camera front and back depth of field according to claim 1, characterized in that: In step 7, when the program crashes due to internal or external factors and the test product lacks test results, the corresponding test product will be judged as an NG product to avoid quality loss, and a test end signal will be generated at the same time.

Images