A method and apparatus for automated detection of paint panels

Abstract

The present application provides a kind of paint test panel automatic detection method and equipment, the method includes the following steps: test panel is extracted from bin by robot control material taking assembly, and test panel is positioned in test table;According to test panel type, difference detection process is executed, detection process includes: if test panel is standard sample, then switch detection component according to preset order to carry out film thickness detection, gloss detection, color detection and particle detection;If test panel is daily board, then the basic two-dimensional code pre-set on the back of daily board is identified by scanning code component to analyze detection item combination, corresponding detection component is dynamically scheduled based on analysis result to execute detection;Based on detection result, unique two-dimensional code is generated, and unique two-dimensional code is sprayed to the back of test panel by code spraying component, and test panel is sorted to corresponding discharge bin according to detection result.The present application can improve detection precision, realize quality data whole process traceability.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of paint test board detection, and in particular to a paint test board automatic detection method and device. BACKGROUND

[0002] In the field of paint test board quality detection, the traditional manual visual inspection method has significant defects: the tester needs to hold a color card, a gloss meter and other equipment to test each sample, and the detection of a single test board takes more than 5 minutes. The difference in the determination of stone peeling area, salt spray corrosion density and other indicators by different personnel can reach more than 20% (such as the determination difference between 3% and 5% of the peeling area), resulting in low data reliability. At the same time, the paper record method causes the detection results to be disconnected from the physical identity of the test board, and cannot realize quality traceability (such as being unable to associate the salt spray corrosion history data of a batch of test boards).

[0003] To alleviate the efficiency and consistency problems caused by manual work, semi-automatic or single-function automatic detection schemes have appeared in the industry: a gloss meter, a color difference meter or a simple visual module are respectively deployed in a fixed station, and a code scanning gun is used to record the test board number. However, these existing automatic improvement schemes still have the following systematic deficiencies:

[0004] Single function: the device such as the gloss meter or the color difference meter only supports single detection, and the standard sample detection and the batch detection of daily samples need to be performed separately, resulting in low device utilization;

[0005] Dynamic scheduling is missing: although the semi-automatic system introduces code scanning recognition, the switching of detection components relies on manual instructions, which takes a long time and cannot optimize the sequence according to the task priority (such as preferentially scheduling the detection items with short time consumption);

[0006] Data island: the detection results are not bound to the physical test board, and are not connected to the cloud quality system, resulting in a broken data traceability chain.

[0007] In view of this, the present application provides a paint test board automatic detection method and device which integrates robot dynamic scheduling, multi-index collaborative detection and quality data closed loop, and can effectively improve the detection efficiency. SUMMARY

[0008] To solve the problems in the current paint test board detection, such as low device utilization caused by isolated function modules, insufficient detection efficiency caused by missing dynamic scheduling, and broken traceability closed loop caused by disconnection of quality data and physical carrier, the present application provides a paint test board automatic detection method and device to solve the above technical defects.

[0009] In a first aspect, the present application provides a paint test board automatic detection method, comprising the following steps:

[0010] S1, extracting the test board from the material bin by the robot-controlled material taking assembly and positioning the test board on the test table;

[0011] S2, performing a differentiated detection process according to the type of test board, the detection process including:

[0012] If the test board is a standard sample, switching the detection components in a preset order to perform film thickness detection, gloss detection, color detection, and particle detection;

[0013] If the test board is a daily board, identifying the basic two-dimensional code pre-set on the back of the daily board through the code scanning assembly to analyze the detection item combination, and dynamically scheduling the corresponding detection components to perform at least one of the following detections based on the analysis result:

[0014] a. Film thickness detection, gloss detection, color detection, and particle detection;

[0015] b. Stone impact defect quantification detection: acquiring the surface image of the daily board through the CCD assembly, using the U-Net neural network to segment the peeling area contour, and calculating the area ratio;

[0016] c. Adhesion defect analysis: extracting the Fourier descriptor features of the coating peeling edge of the daily board and comparing the similarity with the reference template;

[0017] d. Salt spray corrosion grade determination: counting the number of corrosion points in the unit area of the daily board and calculating the distribution density;

[0018] S3, generating a unique two-dimensional code based on the detection results, printing the unique two-dimensional code on the back of the test board through the inkjet assembly, and sorting the test board to the corresponding discharge bin according to the detection results.

[0019] Further preferably, in step S2, when performing stone impact defect quantification detection, the following sub-steps are included:

[0020] b1, the robot moves the CCD assembly above the test table, triggering the coaxial light source to start the dark field illumination mode;

[0021] b2, acquiring the surface image of the daily board and transmitting it to the upper computer;

[0022] b3, the upper computer calls the pre-trained U-Net model to segment the peeling area of the daily board surface image, and generates a binary contour map;

[0023] b4, calculating the peeling area ratio based on the binary contour map, and determining the defect grade according to the ratio value to obtain the detection result.

[0024] Further preferably, in step S2, when performing adhesion defect analysis, the following sub-steps are included:

[0025] c1. The robot moves the CCD assembly above the test table, switches the CCD assembly to the bright field illumination mode, and collects the peeling area image;

[0026] c2. The Canny operator is used to perform edge detection on the peeling area image, and the gradient threshold is set to obtain the continuous coordinate sequence of the coating peeling edge;

[0027] c3. Discrete Fourier transform is performed on the continuous coordinate sequence to generate a Fourier descriptor sequence;

[0028] c4. Based on the Fourier descriptor sequence, the Euclidean distance from the reference template is calculated, and the adhesion failure is determined by judging whether the Euclidean distance is greater than or equal to the distance threshold, to obtain the detection result.

[0029] Further preferably, in step S2, when performing salt spray corrosion grade determination, the following sub-steps are included:

[0030] d1. The robot moves the CCD assembly above the test table, switches the CCD assembly to the bright field illumination mode, and collects multiple sets of surface images at a preset interval along the length direction of the daily plate;

[0031] d2. Perform morphological opening operation on each set of surface images to separate the adherent area and identify the independent corrosion point profile that satisfies the area greater than the preset area threshold;

[0032] d3. Based on the independent corrosion point profile, the number of effective corrosion points in the standard detection area is counted, and then the distribution density is calculated;

[0033] d4. According to the distribution density, the corrosion grade is divided, and the detection result is obtained.

[0034] Preferably, in step S1, the robot controls the material taking assembly to extract the test plate from the material bin, and positions the test plate on the test table, specifically including the following sub-steps:

[0035] S11. The robot end installs a vacuum suction disc type material taking assembly through an electrically coupled quick change interface;

[0036] S12. Select the standard sample bin or daily plate bin according to the test plate type;

[0037] S13. Start the pneumatic jacking mechanism at the bottom of the standard sample bin or daily plate bin, vertically lift the bottom layer of test plates in the bin to a set height, and make the top layer of test plates in the taking position;

[0038] S14. After the material taking assembly adsorbs the top layer of test plates, the robot moves out of the bin along the horizontal direction;

[0039] S15. Accurately position the test plate at the right angle positioning edge of the test table.

[0040] Preferably, in step S2, before detection, an environmental parameter compensation step is further included:

[0041] The ambient environmental parameters around the test bench are monitored in real time by a temperature and humidity sensor;

[0042] When the ambient temperature deviates from the standard value, linear compensation is performed on the film thickness measurement value according to the thermal expansion coefficient α of the substrate, and the compensation formula is: Δd = d × α × ΔT, where Δd is the thickness compensation amount, d is the measured film thickness value, α is the thermal expansion coefficient of the substrate, and ΔT is the difference between the current ambient temperature and the standard value;

[0043] When the ambient humidity exceeds the preset relative humidity value, the dry gas flow intensity in the gloss meter probe is increased;

[0044] When the ambient illuminance fluctuation is greater than the preset illuminance, the light shield of the color difference meter is triggered and the built-in calibration light source is started.

[0045] Preferably, in step S2, the detection process of the standard sample includes:

[0046] S211, the robot switches to the gloss meter assembly through the quick-change interface, and performs surface gloss measurement of the test plate based on the preset angle parameters;

[0047] S212, the robot switches to the color difference meter assembly, collects the colorimetric data of the test plate in the CIE L*a*b* color space, and compares the result with the configurable color difference threshold value;

[0048] S213, the robot switches to the CCD assembly, turns on the coaxial light source and collects the surface image, identifies the surface particle contour of the surface image through image segmentation algorithm, calculates the particle size distribution density, and marks the particle defects when the particle size exceeds the threshold value or the distribution density exceeds the standard;

[0049] S214, the robot switches to the film thickness meter assembly, measures the thickness of the test plate surface coating, and marks the film thickness abnormality if the film thickness value deviates from the preset standard value.

[0050] Preferably, in step S2, the corresponding detection assembly is dynamically scheduled based on the analysis result, including the following sub-steps:

[0051] S221, the code scanning assembly reads the basic two-dimensional code on the back of the daily plate, and extracts the daily plate ID and detection item code;

[0052] S222, the detection sequence instruction is generated by querying the preset rule library according to the detection item code; the detection sequence instruction includes the detection component type, the detection sequence and the parameter configuration;

[0053] S223, the robot switches the detection assembly through the quick-change interface according to the detection sequence instruction, and monitors the assembly switching state in real time, and if the switching time is greater than the preset time length, an alarm is triggered and a fault code is recorded.

[0054] Preferably, in step S3, a unique two-dimensional code is generated based on the detection result, the unique two-dimensional code is printed on the back of the test plate through the code printing assembly, and the test plate is sorted into the corresponding unloading bin according to the detection result, and the specific steps include the following sub-steps:

[0055] S31, the industrial computer receives the detection result data, binds it with the test plate ID and the detection timestamp, and generates a unique two-dimensional code containing the following fields: test plate material and size, detection item result value, BASF Digilab system traceability link, and encrypted check code;

[0056] S32, the robot carries the test plate to the code printing station, and the UV code printer prints the two-dimensional code on the back of the test plate at an incident angle of 30°-60° under the adjustment of the multi-degree-of-freedom support;

[0057] S33, according to the defect type in the detection result, the test plate is sorted into the standard sample OK bin, the standard sample NG bin or the daily plate bin.

[0058] In a second aspect, the present application provides an automatic detection equipment for paint test plates, which is used to realize any of the above-mentioned automatic detection methods for paint test plates, and comprises:

[0059] A robot, the end of the robot is provided with a vacuum suction type material taking assembly through an electrically coupled quick-change interface;

[0060] A feeding bin, the feeding bin includes a standard sample bin and a daily plate bin, and a pneumatic jacking mechanism is arranged at the bottom of the feeding bin;

[0061] A detection assembly library, the detection assembly library includes a film thickness meter assembly, a gloss meter assembly, a color difference meter assembly, and a CCD assembly, and the CCD assembly is provided with a coaxial light source and a bright field / dark field switching controller;

[0062] A test table, the test table is provided with a servo-driven precision transplanting platform and an adjustable right-angle positioning edge;

[0063] A code scanning module, the code scanning module is positioned at the entrance side of the test table;

[0064] A code printing module, the code printing module is provided with a multi-degree-of-freedom adjustment support, and the nozzle is directed to the exit side of the test table;

[0065] An unloading bin, the unloading bin includes a standard sample OK bin, a standard sample NG bin and a daily plate bin;

[0066] A controller is connected with the robot, the detection component library, the code scanning module and the code spraying module, and the controller is internally provided with the following modules:

[0067] A rule analysis module is used for decoding the basic two-dimensional code of the daily board and generating a detection sequence;

[0068] A dynamic scheduling module is used for controlling the robot to switch the detection components according to the sequence;

[0069] A defect analysis module is used for performing stone peeling segmentation, adhesion profile comparison and salt mist density statistics;

[0070] A data encryption module is used for binding the detection result and the test board ID to generate a unique two-dimensional code.

[0071] Compared with the prior art, the beneficial results of the present application are as follows:

[0072] (1) The detection efficiency is significantly improved: the robot realizes instantaneous switching of the detection components in combination with the quick-change interface, and continuously supplies materials by means of the pneumatic jacking mechanism, so that the single piece beat is significantly compressed, and the continuous production demand is met; the standard sample plate and the daily board share the same test table, avoiding separate operation, and the equipment utilization rate is greatly improved.

[0073] (2) The defect judgment is objective and quantifiable: the stone peeling area is segmented at the pixel level through the deep learning model, the area ratio calculation is accurate, the Fourier descriptor is used for profile comparison of the adhesion peeling edge, and the similarity judgment is reliable; the salt mist corrosion density is based on the corrosion point distribution density, realizes digital grading of the corrosion grade, and can eliminate the subjective error of manual visual inspection.

[0074] (3) Dynamic combination of detection items: the two-dimensional code on the back of the daily board carries the detection item code, the system generates a personalized detection sequence in real time through the rule analysis module, avoids the resource waste caused by the fixed menu, and only needs to extend the code in the rule library to realize the function expansion without hardware modification.

[0075] (4) Quality data is traceable throughout the process: the detection result, the test board identity, the timestamp and the material information are written into the unique two-dimensional code after encryption, and are synchronously uploaded to the quality traceability system to form a "test board-data" closed loop; when an exception occurs, the specific batch and the detection process can be quickly located, and the traceability efficiency is greatly improved.

[0076] (5) Small equipment occupation and flexible layout: the whole machine structure is compact, the quick-change interface integrates multiple detection components at the end of the same robot, and the space required by the traditional rotary table or multi-station sliding table is significantly reduced; the right-angle positioning edge is compatible with the precise transplanting platform for test boards of various specifications, and the mechanical adjustment is not required for changeover, and the adaptability is strong.

[0077] (6) Low operation and maintenance cost, can be quickly replicated: all detection components adopt modularized sub-disk design, convenient to disassemble and assemble; the control system reserves standard interface, when subsequent production line is replicated, only needs to import parameters to deploy in batches, the debugging period is greatly shortened, and the maintenance cost is significantly reduced. BRIEF DESCRIPTION OF DRAWINGS

[0078] Other features, objects, and advantages of the application will become more apparent from the following detailed description of non-limiting embodiments, read in conjunction with the accompanying drawings:

[0079] Figure 1 is a flow chart of the paint test plate automatic detection method according to the application;

[0080] Figure 2 is a schematic diagram of an adhesion detection positioning calibration plate according to the application;

[0081] Figure 3 is a schematic diagram of a coating peeling feature ratio according to the application;

[0082] Figure 4 is a schematic diagram of local corrosion quantitative analysis according to the application;

[0083] Figure 5 is a schematic diagram of multi-dimensional corrosion evaluation comparison according to the application;

[0084] Figure 6 is a physical layout top view of the automatic detection equipment according to the application;

[0085] Figure 7 shows a material taking assembly structure schematic diagram of the application;

[0086] Figure 8 shows a gloss meter assembly structure schematic diagram of the application;

[0087] Figure 9 shows a color difference meter assembly structure schematic diagram of the application;

[0088] Figure 10 shows a CCD assembly structure schematic diagram of the application;

[0089] Figure 11 shows a film thickness meter assembly structure schematic diagram of the application.

[0090] Fig. 1, robot, 2, material taking assembly, 21, instrument mounting mechanism of material taking assembly, 22, quick-change tool disc of material taking assembly, 23, tool placing mechanism of material taking assembly, 24, suction cup, 31, standard sample plate warehouse, 32, daily plate warehouse, 41, gloss meter assembly, 411, instrument mounting mechanism of gloss meter assembly, 412, quick-change tool disc of gloss meter assembly, 413, tool placing mechanism of gloss meter assembly, 42, color difference meter assembly, 421, instrument mounting mechanism of color difference meter assembly, 422, quick-change tool disc of color difference meter assembly, 423, tool placing mechanism of color difference meter assembly, 43, CCD assembly, 431, instrument mounting mechanism of CCD assembly, 432, quick-change tool disc of CCD assembly, 433, tool placing mechanism of CCD assembly, 434, lens, 435, coaxial light source, 436, camera, 44, film thickness meter assembly, 441, instrument mounting mechanism of film thickness meter assembly, 442, quick-change tool disc of film thickness meter assembly, 443, tool placing mechanism of film thickness meter assembly, 5, test table, 6, code scanning module, 7, code spraying module, 81, standard sample OK warehouse, 82, standard sample NG warehouse, 83, daily plate warehouse. DETAILED DESCRIPTION

[0091] The application will be further described below in detail with reference to the drawings and embodiments. It can be understood that the specific embodiments described herein are only used to explain the related application, and not to limit the application. In addition, it should be noted that only the parts related to the application are shown in the drawings for ease of description.

[0092] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the drawings and embodiments.

[0093] The present application provides a paint test plate automatic detection method, Figure 1 The flow chart of the paint test plate automatic detection method of the present application is shown, as Figure 1 The method comprises the following steps:

[0094] S1, a robot controls a material taking assembly to extract a paint test plate from a warehouse, and positions the test plate on a test table, preferably the paint test plate is a paint test plate, and specifically comprises the following sub-steps:

[0095] S11, a vacuum suction cup type material taking assembly is installed at the end of the robot through an electrically coupled quick-change interface;

[0096] S12, a standard sample plate warehouse or a daily plate warehouse is selected according to the type of the test plate;

[0097] S13, start the pneumatic jacking mechanism of the standard sample bin or the daily sample bin, vertically lift the bottom sample to a certain height, and make the top sample be at the sample taking position;

[0098] S14, after the sample taking assembly adsorbs the top sample, the robot moves out of the bin along the horizontal direction;

[0099] S15, accurately position the sample on the right-angle positioning edge of the test table. During positioning, ensure that the back of the sample is upward so that the code scanning assembly can read the two-dimensional code.

[0100] In a specific embodiment, the six-axis industrial robot first installs a vacuum chuck type sample taking assembly through the electrical coupling quick change interface of the end flange. The main disc of the quick change interface is fixed to the sixth axis of the robot, and the secondary disc is integrated on the back of the sample taking assembly, so that the air circuit and signal can be connected at one time without additional manual connection.

[0101] When the system identifies that the current batch needs to detect the standard sample (master sample), the robot automatically turns to the standard sample bin; if it is a daily sample, it turns to the daily sample bin. Both bins adopt a structure of “right-angle double-side positioning + bottom pneumatic jacking”: a jacking mechanism driven by a cylinder is arranged below the bottom plate of the bin, which can push the whole stack of samples upward, so that the uppermost sample is always at the same sample taking height, ensuring the repeated positioning accuracy of the robot. The side wall of the bin is provided with adjustable blocking strips, which are compatible with aluminum plates or tin plates of various specifications such as 150×100 mm, 200×100 mm and 90×190 mm, and prevent the samples from moving horizontally during jacking through right-angle two-side positioning.

[0102] After the robot reaches above the bin, the sample taking assembly is lowered, the vacuum chuck contacts the upper surface of the sample, and the vacuum generator instantly establishes negative pressure to firmly adsorb the sample. Then the robot moves out of the bin along the horizontal direction smoothly to avoid collision with the wall. The whole sample taking process is completed under the interlocked state of the safety door, and if the door is accidentally opened, the robot will immediately execute emergency stop to prevent personal injury.

[0103] The robot moves to the test table entrance with the sample. The test table is composed of a precision transplanting platform driven by a servo motor and an adjustable right-angle positioning edge: the surface of the transplanting platform is embedded with low-friction balls to reduce the sliding resistance of the sample; the right-angle positioning edge realizes two-stage positioning of “forward pushing + side pushing” through a miniature cylinder, so as to ensure that the long side and the short side of the sample are simultaneously close to the reference surface. After positioning is completed, the transplanting platform is lowered to the detection plane to enter the subsequent detection process.

[0104] Through the above steps, the robot can automatically take the standard sample and the daily sample, accurately position them in the same beat without manual intervention, and lay a stable foundation for subsequent differentiated detection.

[0105] With continued reference to Figure 1The paint test board automatic detection method provided by the application further includes the following steps:

[0106] S2, performing a differentiated detection process according to the type of the test board, the detection process including:

[0107] If the test board is a standard sample, switch the detection components in a preset order to perform film thickness detection, gloss detection, color detection and particle detection;

[0108] If the test board is a daily board, identify the basic two-dimensional code preset on the back of the daily board through the code scanning component to analyze the detection item combination, and based on the analysis result, dynamically schedule the corresponding detection components to perform at least one of the following detections:

[0109] a, gloss detection, color detection, particle detection and film thickness detection;

[0110] b, stone impact defect quantitative detection: the CCD component collects the surface image of the daily board, and the U-Net neural network is used to segment the peeling area contour and calculate the area ratio;

[0111] c, adhesion defect analysis: extracting the Fourier descriptor features of the peeling edge of the daily board coating, and comparing the similarity with the reference template;

[0112] d, salt spray corrosion grade determination: counting the number of corrosion points in the unit area of the daily board and calculating the distribution density.

[0113] Before performing the detection in step S2, an environmental parameter compensation step is further included:

[0114] The ambient environmental parameters around the test bench are monitored in real time through a temperature and humidity sensor;

[0115] When the ambient temperature deviates from the standard value (for example, 23℃), the film thickness measurement value is linearly compensated according to the thermal expansion coefficient α of the substrate, and the compensation formula is: Δd=d×α×ΔT, wherein Δd is the thickness compensation amount, d is the measured film thickness value, α is the thermal expansion coefficient of the substrate, and ΔT is the difference between the current ambient temperature and the standard value;

[0116] When the ambient humidity exceeds a preset relative humidity value (for example, 60%RH), the dry gas flow intensity in the gloss meter probe is increased, such as to 1.5 L / min;

[0117] When the ambient illuminance fluctuation is greater than a preset illuminance (for example, ±50 lux), the color difference instrument light shield is triggered and the built-in calibration light source is started;

[0118] All compensation operation logs are written into the final report with the detection results.

[0119] In this embodiment, when performing gloss detection, color detection, particle detection and film thickness detection (the execution process of the standard sample and the daily board is completely consistent, only for the daily board, the detection sequence needs to be inserted based on dynamic scheduling), the following sub-steps are included:

[0120] a1, gloss detection execution: the robot switches to the BYK gloss meter assembly through the quick-change interface; the gloss meter probe is vertically aligned with the surface of the test board (such as a distance of 50±0.1 mm), and the gloss at an incident angle of 60° (or other preset angles such as 20°, 85°, etc.) is measured according to the GB / T 9754 standard:

[0121] (1) Perform three repeated measurements (with an interval of 0.5 s), and take the average value as the final value;

[0122] (2) If the fluctuation value is greater than 2 GU, trigger the automatic re-measurement mechanism (up to 3 re-measurements);

[0123] (3) The measurement value is uploaded to the industrial computer database in real time.

[0124] a2, color detection execution: the robot switches to the BYK colorimeter assembly; the colorimetric data of the test board surface is collected under the D65 standard light source, and the CIE L*a*b* values (accuracy ΔE≤0.1) are output;

[0125] Compare with the configurable color difference threshold to determine the result, which can be specifically compared with the preset standard color plate for color difference ΔE. If ΔE≤1.0, it is determined to be qualified; if 1.0<ΔE≤2.0, it is marked as a slight deviation in color; if ΔE>2.0, it is marked as a serious color difference (marked as NG).

[0126] a3, particle detection preparation: the robot switches to the CCD assembly and turns on the coaxial light source bright field mode (illuminance 1500±100 lux); adjust the lens focal length to make the imaging resolution reach 0.05 mm / pixel, and collect the surface image.

[0127] a4, particle defect quantification: separate the particle area by morphological top-hat transformation; calculate the particle size distribution density: identify the particle outline with a diameter greater than 50 μm; count the number of particles per unit area (1 cm 2 ); if the particle density is greater than 5 per cm 2 or there is a particle with a diameter greater than 100 μm, it is determined to be a particle defect.

[0128] a5, film thickness detection: the robot switches to the film thickness gauge assembly and performs five-point positioning measurement (center point + four corner points, distance ≥20 mm) on the surface of the test board; the coating thickness is measured by a magnetic induction / vortex dual-mode probe contact measurement, and the single-point measurement pressure is 0.5±0.1 N;

[0129] If the film thickness value deviates from the preset standard value, the deviation rate δ is calculated according to the following formula, and the following is determined:

[0130]

[0131] If the deviation rate δ of any point is greater than 10%, the film thickness is marked as abnormal;

[0132] If the deviation rate δ of all points is less than or equal to 5%, it is determined to be qualified.

[0133] When performing stone impact defect quantitative detection, the following sub-steps are included:

[0134] b1, the robot moves the CCD assembly to the center position above the test table, 50±0.1mm away from the surface of the test plate, triggers the coaxial light source to start the dark field illumination mode (wavelength 400-700nm);

[0135] b2, collect the daily plate surface image and transmit it to the upper computer;

[0136] b3, the upper computer calls the pre-trained U-Net model to segment the spalling area of the daily plate surface image, and generates a binary contour map;

[0137] b4, calculate the spalling area ratio based on the binary contour map, and determine the defect grade according to the ratio value to obtain the detection result, the ratio formula is as follows:

[0138]

[0139] Wherein, R represents the ratio value, if the ratio value R is greater than or equal to 5%, it is determined that the daily plate has serious defects.

[0140] Figure 2 The schematic diagram of the adhesion detection positioning calibration plate of the present application is shown, Figure 3 The schematic diagram of the coating peeling feature ratio of the present application is shown; in combination with reference Figure 1 、 Figure 2 and Figure 3 When performing adhesion defect analysis, the following sub-steps are included:

[0141] c1, the robot moves the CCD assembly to the test table above, aligns the test plate mark area (such as the green box position in Figure 2 ), switches the CCD assembly to the bright field illumination mode (5500K color temperature), and the resolution is 0.05mm / pixel peeling area image ( Figure 3 ruler parameters);

[0142] c2, the Canny operator is used to perform edge detection on the peeling area image, and the gradient threshold value (the preferred gradient threshold value is 50) is set to obtain the continuous coordinate sequence of the coating peeling edge;

[0143] c3. Discrete Fourier transform is performed on the continuous coordinate sequence to generate a Fourier descriptor sequence;

[0144] c4. Euclidean distance is calculated based on the Fourier descriptor sequence and the reference template, and whether the adhesion fails is determined by judging whether the Euclidean distance is greater than or equal to a distance threshold, to obtain a detection result.

[0145] Figure 4 A local corrosion quantification analysis schematic diagram of the present application is shown, Figure 5 A multi-dimensional corrosion evaluation comparison schematic diagram of the present application is shown, in combination with reference Figure 1 、 Figure 4 and Figure 5 When performing salt spray corrosion grade determination, the following sub-steps are included:

[0146] d1. The robot moves the CCD assembly above the test table, switches the CCD assembly to the bright field illumination mode, and acquires multiple groups of surface images at a preset interval (for example, 10±0.5mm interval) along the length direction of the daily plate (reference Figure 5 Multi-region comparison);

[0147] d2. Perform morphological opening operation on each group of surface images to separate the adherent regions and identify independent corrosion point profiles that satisfy a preset area threshold (reference Figure 5 Green box, actual optional 25cm 2 );

[0148] d3. Based on the independent corrosion point profiles, the number of effective corrosion points in the standard detection area is counted, and then the distribution density ρ is calculated;

[0149] d4. According to the distribution density ρ, the corrosion grade is divided, and the detection result is obtained. For example, if ρ<5, it is determined that the daily plate is G0 grade, and it needs to be sorted to the standard sample OK bin subsequently; if 5≤ρ≤20, it is determined that the daily plate is G1 grade, and it needs to be sorted to the daily plate bin subsequently; if ρ is greater than 20, it is determined that the daily plate is G2 grade, and it needs to be sorted to the standard sample NG bin subsequently.

[0150] In specific embodiments, the detection process of the standard sample includes:

[0151] S211. The robot switches to the gloss meter assembly through the quick-change interface, and performs surface gloss measurement of the test plate based on preset angle parameters (for example, the preset angle parameters are 20°, 60° or 85°);

[0152] S212. The robot switches to the color difference meter assembly, acquires the colorimetric data of the test plate in the CIE L*a*b* color space, and compares with a configurable color difference threshold to determine the result;

[0153] S213, the robot switches to the CCD component, turns on the coaxial light source and collects the surface image, identifies the surface particle contour of the surface image through the image segmentation algorithm, calculates the particle size distribution density, and marks the particle defect when the particle size exceeds the threshold or the distribution density exceeds the standard;

[0154] S214, the robot switches to the film thickness gauge component, measures the thickness of the surface coating of the test plate through the magnetic induction / eddy current dual-mode probe, obtains the film thickness distribution data, and marks the film thickness abnormality if the film thickness value deviates from the preset standard value.

[0155] The gloss meter component, the color difference meter component, and the film thickness gauge component are connected to the data line through the electrical pins of the quick-change interface, and a hierarchical communication protocol is adopted: the gloss meter adopts the RS485-Modbus protocol (transmission delay ≤10 ms); the color difference meter adopts the Modbus-TCP protocol (transmission delay ≤5 ms); and the film thickness gauge component adopts the EtherCAT protocol (transmission delay ≤1 ms).

[0156] Based on the analysis result, the corresponding detection component is dynamically scheduled, including the following sub-steps:

[0157] S221, the code scanning component reads the basic two-dimensional code on the back of the daily board, and extracts the daily board ID and detection item code;

[0158] S222, the preset rule library is queried according to the detection item code to generate a detection sequence instruction; the detection sequence instruction includes the detection component type, the detection sequence, and the parameter configuration;

[0159] S223, the robot switches the detection components in sequence through the quick-change interface according to the detection sequence instruction, and monitors the component switching state in real time; if the switching time is greater than the preset time length (for example, 1 second), an alarm is triggered and a fault code is recorded.

[0160] In the automatic detection method of the paint test plate, an automatic calibration and state monitoring mechanism for the detection components is further included: before each detection task starts, the robot controls the standard calibration plate to be placed on the test table, and each detection component is switched in sequence to perform a self-calibration process; wherein the gloss meter component performs deviation correction by measuring the known gloss value (such as 100 GU) of the calibration plate, the color difference meter component measures the L*a*b* value of the standard color plate and calculates the color difference offset, and if ΔE ≥ 0.3, a soft calibration is triggered; the CCD component performs distortion correction and pixel calibration by shooting a standard grid image, ensuring that the imaging resolution is stable at 0.05 mm / pixel; the film thickness gauge component performs zero-point calibration by measuring a standard piece with a known thickness (such as 100 μm), and if the measurement deviation exceeds ±1 μm, the probe pressure is automatically adjusted; all calibration data are uploaded to the industrial computer in real time and recorded to a log, and if the calibration fails, the detection is paused and an alarm is triggered, ensuring that the detection data is reliable and traceable throughout the process.

[0161] With reference to the above Figure 1 The paint test plate automatic detection method provided by the present application further comprises the following steps:

[0162] S3, based on the detection result, a unique two-dimensional code is generated, the unique two-dimensional code is sprayed to the back of the test plate through a code spraying component, and the test plate is sorted into a corresponding unloading bin according to the detection result, and specifically comprises the following sub-steps:

[0163] S31, the industrial computer receives the detection result data, binds it with the test plate ID and the detection timestamp, and generates a unique two-dimensional code containing the following fields: test plate material and size, detection item result value, BASF Digital Laboratory Platform (BASF Digital Laboratory Platform) traceability link, and encrypted check code;

[0164] S32, the robot carries the test plate to the code spraying station, and the UV code spraying machine sprays the two-dimensional code on the back of the test plate at an incident angle of 30°-60° under the adjustment of the multi-degree-of-freedom support;

[0165] S33, according to the defect type in the detection result, the test plate is sorted into a standard sample OK bin, a standard sample NG bin or a daily plate bin.

[0166] The three bins all adopt a bottom detachable drawer structure, and a single stack can stack not less than 200 test plates; after the drawer is in place, the photoelectric sensor confirms the full bin state and prompts manual or AGV to change the bin. The sorting path is programmed offline by the robot to ensure no collision with the bin entrance; at the same time, the equipment safety interlocking door remains locked until the robot completely exits the bin area.

[0167] In the paint test plate automatic detection method, an intelligent learning and optimization mechanism based on edge computing and cloud cooperation is further included: the system automatically starts a model optimization period after completing 1000 detections, and the specific steps are as follows: the industrial computer extracts historical detection data (including image, spectrum data and measurement value) and transmits it to the cloud training platform for encryption; the cloud training platform updates the defect recognition model using an incremental learning algorithm: for stone impact defect segmentation, the U-Net model convolution kernel weight is optimized using new samples to improve the recognition sensitivity of small area peeling (<0.5%); for adhesion analysis, the Fourier descriptor template library is expanded by adding peeling contour samples, and a dynamic Euclidean distance threshold (initial threshold ±3σ adaptive adjustment) is used; the optimized model is signed and distributed to each terminal device, and the robot receives the new model and automatically enters the verification mode: 3 test plates in the standard sample library are randomly selected for comparison test, if the consistency of the new model detection result and the traditional method is more than 98%, the new model is applied, otherwise, the last version is rolled back and a technical intervention warning is triggered, realizing the continuous self-evolution of the detection algorithm.

[0168] In a second aspect, the present application provides an automatic detection equipment for paint test board, which is used to implement any of the above methods, Figure 6 The automatic detection equipment of the present application is shown in the top view of physical layout as Figure 6 The automatic detection equipment comprises:

[0169] A robot 1, which is equipped with a vacuum suction type material taking assembly 2 at the end of the robot through an electrically coupled quick-change interface;

[0170] A loading bin, which comprises a standard sample board bin 31 and a daily board bin 32, and is provided with a pneumatic jacking mechanism at the bottom;

[0171] A detection assembly library, which comprises a gloss meter assembly 41, a color difference meter assembly 42, a CCD assembly 43 and a film thickness meter assembly 44, wherein the CCD assembly 43 is equipped with a coaxial light source and a bright field / dark field switching controller;

[0172] A test table 5, which is provided with a servo-driven precision transplanting platform and an adjustable right-angle positioning edge;

[0173] A code scanning module 6, which is positioned at the entrance side of the test table;

[0174] A code printing module 7, which is provided with a multi-degree-of-freedom adjusting support, and the nozzle is directed to the exit side of the test table;

[0175] A discharging bin, which comprises a standard sample board OK bin 81, a standard sample board NG bin 82 and a daily board bin 83;

[0176] A controller, which is connected with the robot 1, the detection assembly library, the code scanning module 6 and the code printing module 7, and is provided with the following modules:

[0177] A rule analysis module, which is used to decode the basic two-dimensional code of the daily board and generate a detection sequence;

[0178] A dynamic scheduling module, which is used to control the robot to switch the detection assembly according to the sequence;

[0179] A defect analysis module, which is used to perform stone impact spalling segmentation, adhesion profile comparison and salt spray density statistics;

[0180] A data encryption module, which is used to bind the detection result and the test board ID to generate a unique two-dimensional code.

[0181] The coating test plate automatic detection equipment provided by the application is also provided with a real-time quality closed loop control mechanism based on multi-sensor data fusion: when the result of any detection item exceeds the threshold value, the system automatically triggers a re-inspection protocol - the robot moves the test plate to a re-inspection station, and a high-resolution CCD component (pixel accuracy 0.01mm) performs local area multi-spectral scanning (including visible light and near-infrared waveband), while fusing the multi-point redundant measurement data of the film thickness gauge (for example, increased to 9-point grid measurement); the industrial computer calculates the initial inspection and re-inspection data by time series analysis algorithm (for example, initial inspection weight 0.4, re-inspection weight 0.6), if the fused data still exceeds the limit, a quality abnormality report is immediately generated and synchronized to the MES system, and the detection of subsequent test plates of the same batch is automatically suspended; at the same time, the system dynamically adjusts the detection parameters: for the color difference abnormality batch, the color difference gauge measurement points are automatically increased from 3 to 5; for the batch with high frequency of particle defects, the ultra-high resolution mode (0.02mm / pixel) of the CCD component is started and Gaussian filter noise reduction processing is performed, realizing adaptive optimization of detection accuracy and efficiency.

[0182] In addition, the coating test plate automatic detection equipment provided by the application is also integrated with temperature and humidity and environmental light sensors for real-time monitoring of the environmental parameters around the test bench. Before high-precision optical detection (such as gloss detection, color detection), the industrial computer calls the preset compensation algorithm model: when the environmental temperature deviates from the standard 23℃, the film thickness measurement value is linearly compensated according to the thermal expansion coefficient of the material; when the environmental humidity exceeds 60%RH, the dry air flow intensity in the gloss meter probe is automatically increased to 1.5L / min to prevent water vapor condensation from affecting the measurement; when the environmental illumination fluctuation is greater than ±50lux, the color difference meter automatically triggers the light shield and starts the built-in calibration light source to ensure that the colorimetric data collection is not disturbed. All compensation operation logs are written into the final detection report together with the environmental data, ensuring that the detection conditions are traceable throughout the process.

[0183] Figure 7 The structure of the material taking assembly of the application is shown in the figure, Figure 7 As shown in the figure, the material taking assembly 2 comprises: an instrument mounting mechanism 21 of the material taking assembly, a quick-change tool disc 22 of the material taking assembly, a tool placing mechanism 23 of the material taking assembly and a suction disc 24. The instrument mounting mechanism 21 of the material taking assembly is fixed to the robot end flange through high-strength bolts, the quick-change tool disc 22 of the material taking assembly is an electromagnetic locking secondary disc, and the coupling response time with the robot main disc is ≤0.1 second, and the surface is engraved with a radial foolproof groove. The tool placing mechanism 23 of the material taking assembly contains a bidirectional spring buffer (stroke 10mm), and the vertical direction shock absorption efficiency is >85% when the assembly is homing, and the horizontal anti-deviation accuracy is ±0.1mm. The suction disc 24 contains two silica gel suction discs, and the suction disc aperture is 3mm. The material taking assembly is built-in with a pressure sensor (range 0-1MPa), which can automatically adjust the negative pressure according to the size of the test plate.

[0184] Figure 8 The structure diagram of the gloss meter assembly of the present application is shown in Fig. 1. Figure 8 As shown in Fig. 1, the gloss meter assembly 41 integrates a quick-change interface system, which includes a machine installation mechanism 411 of the gloss meter assembly, a quick-change tool disc 412 of the gloss meter assembly, and a tool placement mechanism 413 of the gloss meter assembly. The machine installation mechanism 411 of the gloss meter assembly is fixed to the body of the gloss meter through high-rigidity support columns, with a positioning repeatability of ±0.05 mm. The quick-change tool disc 412 of the gloss meter assembly is electromagnetically coupled to the main disc of the robot as a secondary disc, with a docking response time of ≤0.1 s. The tool placement mechanism 413 of the gloss meter assembly is a positioning clamping groove with spring buffering, which buffers the impact force when the assembly is recovered.

[0185] Figure 9 The structure diagram of the color difference meter assembly of the present application is shown in Fig. 2. Figure 9 As shown in Fig. 2, the color difference meter assembly 42 integrates a quick-change interface system, which includes a machine installation mechanism 421 of the color difference meter assembly, a quick-change tool disc 422 of the color difference meter assembly, and a tool placement mechanism 423 of the color difference meter assembly. The machine installation mechanism 421 of the color difference meter assembly is fixed to the body of the color difference meter through high-rigidity support columns, with a positioning repeatability of ≤±0.05 mm. The quick-change tool disc 422 of the color difference meter assembly is electromagnetically coupled to the main disc of the robot through an electromagnetic locking device, with a response time of ≤0.1 s and a current signal confirmation state (5-24 VDC). The tool placement mechanism 423 of the color difference meter assembly is a positioning clamping groove with spring buffering, with a maximum buffering stroke of 10 mm.

[0186] Figure 10 The structure diagram of the CCD assembly of the present application is shown in Fig. 3. Figure 10 As shown in Fig. 3, the CCD assembly 43 includes a machine installation mechanism 431 of the CCD assembly, a quick-change tool disc 432 of the CCD assembly, a tool placement mechanism 433 of the CCD assembly, a lens 434, a coaxial light source 435, and a camera 436. The camera 436 is vertically installed in the middle of the top plate and is electromagnetically coupled to the main disc of the robot through the quick-change tool disc 432 of the CCD assembly, with a response time of ≤0.1 s. The machine installation mechanism 431 of the CCD assembly adopts a high-rigidity aluminum alloy frame to ensure a camera positioning repeatability of ±0.05 mm. The lens 434 is equipped with a telecentric optical system, with a working distance of 50±0.1 mm. The coaxial light source 435 is located in front of the lens and covers a wavelength range of 400-700 nm, with an illuminance of 2000±100 lux. The tool placement mechanism 433 of the CCD assembly is internally provided with a silicone buffer pad with a Shore hardness of 60A, with a maximum buffering stroke of 8 mm and a shock absorption efficiency of >90% when the assembly is homed.

[0187] Figure 11 The structure diagram of the film thickness meter assembly of the present application is shown in Fig. 4. Figure 11As shown, the film thickness gauge assembly 44 comprises: a film thickness gauge assembly instrument mounting mechanism 441, a film thickness gauge assembly quick-change tool disc 442, and a film thickness gauge assembly tool placement mechanism 443. The film thickness gauge assembly instrument mounting mechanism 441 fixes the film thickness gauge body (model: ElektroPhysik MiniTest 740) through a high-rigidity stainless steel support, so that the repeatability of the positioning accuracy is ±0.05 mm. The film thickness gauge body is internally provided with a piezoelectric sensor, which can adjust the probe contact pressure to 0.5±0.1 N in real time. The film thickness gauge assembly quick-change tool disc 442 is electromagnetically coupled with the robot main disc, and the response time is ≤0.1. The film thickness gauge assembly tool placement mechanism 443 is a positioning clamping groove with spring buffering, and the maximum buffering stroke is 10 mm. The film thickness gauge assembly 44 further comprises a dual-mode probe, which can automatically switch between a magnetic induction mode and an eddy current mode, and the switching response is ≤0.2 seconds, and the measurement accuracy is ±1 μm. The film thickness gauge assembly 44 further comprises a cross positioning support to realize five-point measurement distribution (center + four corners, spacing ≥20 mm).

[0188] The above description is merely the preferred embodiments of the present application and the explanation of the technical principles used. It should be understood by those skilled in the art that the scope of the application involved in the present application is not limited to the technical solutions formed by the specific combinations of the above technical features, and should also cover other technical solutions formed by any combinations of the above technical features or their equivalent features without departing from the above inventive concept. For example, the above features can be replaced with the technical features disclosed in the present application (but not limited to) having similar functions to form technical solutions.

Claims

1. A method for automated detection of paint panels, characterized by, The method comprises the following steps: S1, extracting a test board from a material bin by a robot-controlled material taking assembly and positioning the test board on a test table; S2, performing a differentiated detection process according to the type of the test board, the detection process comprising: if the test board is a standard sample, switching detection assemblies in a preset order to perform film thickness detection, gloss detection, color detection and particle detection; if the test board is a daily board, identifying a basic two-dimensional code preset on the back of the daily board by a code scanning assembly to analyze a detection item combination, and dynamically scheduling corresponding detection assemblies to perform the following detections based on the analysis result: a. film thickness detection, gloss detection, color detection and particle detection; b. stone impact defect quantification detection: acquiring a surface image of the daily board by a CCD assembly, segmenting a peeling area contour by a U-Net neural network, and calculating an area proportion; c. adhesion defect analysis: extracting Fourier descriptor features of a coating peeling edge of the daily board, and comparing the similarity with a reference template; d. salt spray corrosion grade determination: counting the number of corrosion points in a unit area of the daily board, and calculating a distribution density; In step S2, when performing adhesion defect analysis, the following sub-steps are included: c1, the robot moves a CCD assembly above the test table, switches the CCD assembly to a bright field illumination mode, and acquires a peeling area image; c2, performing edge detection on the peeling area image by a Canny operator, setting a gradient threshold, and obtaining a continuous coordinate sequence of the coating peeling edge; c3, performing discrete Fourier transform on the continuous coordinate sequence to generate a Fourier descriptor sequence; c4, calculating the Euclidean distance with a reference template based on the Fourier descriptor sequence, and determining whether the adhesion is failed by judging whether the Euclidean distance is greater than or equal to a distance threshold, to obtain a detection result; In step S2, before performing detection, an environmental parameter compensation step is further included: real-time monitoring of environmental parameters around the test table by a temperature and humidity sensor; when the environmental temperature deviates from the standard value, linearly compensating the film thickness measurement value according to the thermal expansion coefficient α of the substrate, and the compensation formula is: Δd=d×α×ΔT, wherein Δd is the thickness compensation amount, d is the measured film thickness value, α is the thermal expansion coefficient of the substrate, and ΔT is the difference between the current environmental temperature and the standard value; when the environmental humidity exceeds a preset relative humidity value, the dry gas flow intensity in the gloss meter probe is increased; when the environmental illuminance fluctuation is greater than a preset illuminance, the light shield of the color difference instrument is triggered and the built-in calibration light source is started; In step S2, the detection process of the standard sample comprises: S211, the robot switches to the gloss meter assembly through the quick-change interface, and performs test board surface gloss measurement based on preset angle parameters; S212, the robot switches to the color difference meter assembly, collects the CIE L a b chrominance data of the color space, and compares the result with a configurable color difference threshold. S213, the robot switches to the CCD assembly, turns on the coaxial light source and acquires a surface image, identifies the surface particle contour of the surface image by an image segmentation algorithm, calculates the particle size distribution density, and marks the particle defect when the particle size exceeds the threshold or the distribution density is out of standard; S214, the robot switches to the film thickness gauge assembly to measure the test board surface coating thickness, and marks the film thickness abnormality if the film thickness value deviates from the preset standard value. S3, generating a unique two-dimensional code based on the detection result, and printing the unique two-dimensional code on the back of the test board through a code printing component, and sorting the test board into a corresponding unloading bin according to the detection result, specifically comprising the following sub-steps: S31, the industrial computer receives the detection result data, binds it with the test board ID and detection timestamp, and generates a unique two-dimensional code containing the following fields: test board material and size, detection item result value, BASF Digilab system traceability link and encrypted check code; S32, the robot carries the test board to the code printing station, and the UV code printer prints a two-dimensional code on the back of the test board at an incident angle of 30°-60° under the adjustment of a multi-degree-of-freedom support; S33, according to the defect type in the detection result, the test board is sorted into a standard sample OK bin, a standard sample NG bin or a daily board bin.

2. The automated paint panel inspection method of claim 1, wherein, In step S2, when performing stone impact defect quantification detection, the following sub-steps are included: b1, the robot moves the CCD component above the test table, and triggers the coaxial light source to start the dark field illumination mode; b2, the surface image of the daily board is collected and transmitted to the upper computer; b3, the upper computer calls the pre-trained U-Net model to segment the peeling area of the daily board surface image, and generates a binary contour map; b4, calculate the peeling area ratio based on the binary contour map, and determine the defect grade according to the ratio value to obtain the detection result.

3. The automated paint panel inspection method of claim 1, wherein, In step S2, when performing salt spray corrosion grade determination, the following sub-steps are included: d1, the robot moves the CCD component above the test table, switches the CCD component to the bright field illumination mode, and collects multiple sets of surface images at a preset interval along the length direction of the daily board; d2, perform morphological opening operation on each set of surface images to separate the adhered areas and identify independent corrosion point contours with an area greater than a preset area threshold; d3, based on the independent corrosion point contours, count the number of effective corrosion points in the standard detection area, and then calculate the distribution density; d4, divide the corrosion grade according to the distribution density to obtain the detection result.

4. The automated paint panel inspection method of claim 1, wherein, In step S1, the robot controls the material taking component to extract the test board from the material bin, and positions the test board on the test table, specifically including the following sub-steps: S11, the robot end installs a vacuum suction disc type material taking component through an electrically coupled quick change interface; S12, select a standard sample bin or a daily board bin according to the test board type; S13, start the pneumatic lifting mechanism at the bottom of the standard sample bin or the daily board bin, vertically lift the bottom layer test board in the bin to a set height, and make the top layer test board in the material taking station; S14, after the material taking component adsorbs the top layer test board, the robot moves out of the bin along the horizontal direction; S15, accurately position the test board at the right angle positioning edge of the test table.

5. The automated paint panel inspection method of claim 1, wherein, In step S2, based on the analysis result, the corresponding detection component is dynamically scheduled, including the following sub-steps: S221, the code scanning component reads the basic two-dimensional code on the back of the daily board, and extracts the daily board ID and detection item code; S222, query a preset rule library according to the detection item code to generate a detection sequence instruction; the detection sequence instruction includes a detection component type, a detection sequence, and parameter configuration; S223, the robot switches the detection components in sequence through the quick-change interface according to the detection sequence instruction, and monitors the component switching state in real time; if the switching time is greater than a preset time length, an alarm is triggered and a fault code is recorded.

6. A paint test panel automated inspection apparatus for implementing the paint test panel automated inspection method according to any one of claims 1 to 5, characterized by, Comprise: a robot, the end of the robot is installed with a vacuum chuck type material taking component through an electrically coupled quick-change interface; a feeding bin, the feeding bin comprises a standard sample bin and a daily board bin, and a pneumatic jacking mechanism is arranged at the bottom of the feeding bin; a detection component library, the detection component library comprises a film thickness instrument component, a gloss instrument component, a color difference instrument component, and a CCD component, the CCD component is equipped with a coaxial light source and a bright field / dark field switching controller; a test table, the test table is provided with a servo driven precision transplanting platform and an adjustable right angle positioning edge; a code scanning module, the code scanning module is positioned at the entrance side of the test table; a code spraying module, the code spraying module is provided with a multi-degree-of-freedom adjusting support, and the nozzle is directed to the exit side of the test table; a discharging bin, the discharging bin comprises a standard sample OK bin, a standard sample NG bin, and a daily board bin; a controller, the controller is connected with the robot, the detection component library, the code scanning module, and the code spraying module, and the controller is built-in the following modules: a rule analysis module, used for decoding a basic two-dimensional code of the daily board and generating a detection sequence; a dynamic scheduling module, used for controlling the robot to switch the detection components in sequence; a defect analysis module, used for performing stone impact peeling segmentation, adhesion force profile comparison, and salt mist density statistics; a data encryption module, used for binding a detection result and a test board ID to generate a unique two-dimensional code.

Images