Visual inspection system and method

Abstract

A visual inspection system and method. The visual inspection system comprises a visual verification device (100), a visual inspection device (200) and a superordinate computer (300), wherein the visual verification device (100) comprises a verification member (10) and a verification member moving device (20), the verification member moving device (20) comprises a movably mounted mounting plate (50), the mounting plate (50) can be located at a verification position on the movement range of the mounting plate (50), the verification position is within an inspection range of the visual inspection device (200), and the verification member (10) is arranged on the mounting plate (50); the visual inspection device (200) is used for acquiring an image of the verification member (10) and sending the image of the verification member (10) to the superordinate computer (300); and the superordinate computer (300) is used for determining a systematic verification result of the visual inspection device (200) on the basis of the image of the verification member (10).

Description

Visual inspection system and method Technical Field

[0001] The present application relates to the technical field of visual inspection. Background Art

[0002] Generally speaking, the calibration of traditional measuring tools is done manually. In addition to being cumbersome and requiring skilled workers, the operation and analysis are time-consuming, resulting in significant productivity losses for wire drawing. Technical issues

[0003] Solve the problem that the existing production line inspection method takes too long to check. Technical Solutions

[0004] In a first aspect, the present application provides a visual inspection system, wherein the visual inspection system comprises:

[0005] A visual inspection device, comprising a verification piece and a verification piece moving device, wherein the verification piece moving device comprises a movably mounted mounting plate, wherein the mounting plate can be positioned in a verification position during its movable travel, wherein the verification position is within the detection range of the visual inspection device, and wherein the verification piece is mounted on the mounting plate;

[0006] A visual inspection device, configured to obtain an image of the verification piece and send the image of the verification piece to a host computer;

[0007] The host computer is used to determine the systematic verification result of the visual inspection device based on the image of the verification part.

[0008] In the technical solution of the embodiment of the present application, when performing system calibration, the calibration parts are concentrated on the mounting plate, and the mounting plate is driven to move the calibration parts, thereby adopting an automatic method to achieve fully automatic calibration, thereby solving the problem that the calibration time is long and the production capacity loss is large due to the influence of the operator's proficiency.

[0009] In some embodiments, the calibration member includes a plurality of size calibration portions with gradient-changing sizes and a color card calibration portion with gradient-changing grayscale values;

[0010] The visual inspection device is further configured to obtain a first image of the size verification portion and a second image of the color card verification portion, and send the first image and the second image to a host computer;

[0011] The host computer is further configured to determine a systematic verification result of the visual inspection device based on the first image and / or the second image.

[0012] In the technical solution of the embodiment of the present application, a plurality of dimension calibration parts with gradient size changes are provided on the calibration piece, which effectively realizes the accuracy calibration within the measuring range of 2D cameras and 3D camera measuring tools, and ensures that the linear offset of the measuring tool is calibrated. The linear offset is the linear error or endpoint linearity, which is the maximum deviation of the straight line composed of the endpoints of the entire measuring range, that is, the deviation between the measured curve and the ideal straight line. At the same time, the accuracy error caused by the distortion of the 2D camera lens can be verified, and on the basis of achieving accuracy calibration, the algorithm calibration is further performed, thereby realizing the systematic calibration of the visual inspection device.

[0013] In some embodiments, the verification piece can be detachably mounted on the mounting plate.

[0014] In the technical solution of the embodiment of the present application, the quick-release locking method of the calibration piece is realized by the detachable part on the calibration piece mounting plate, which is convenient for true value measurement of the two-dimensional or higher precision measurement system.

[0015] In some embodiments, the visual verification device further comprises:

[0016] a support base; and,

[0017] The driving device comprises a fixed part and a movable part which move linearly relative to each other, the fixed part is arranged on the supporting base, and the movable part is connected to the mounting plate.

[0018] In the technical solution of the embodiment of the present application, the movement of the verification piece is achieved by driving, thereby realizing fully automatic verification of the visual inspection device.

[0019] In some embodiments, the visual verification device also includes a sliding guide structure, which includes a sliding rail and a slider that slide together, and the sliding guide structure includes a sliding rail and a slider that slide together, and the sliding rail and the slider are arranged between the support base and the mounting plate.

[0020] In the technical solution of the embodiment of the present application, when the verification piece mounting plate is driven by the driving device, the movement trajectory of the verification piece mounting plate is determined by the sliding device, thereby improving the accuracy of the movement of the verification piece.

[0021] In some embodiments, the visual verification device further includes an adapter, two of which are provided for the sliders, and two corresponding to the sliders;

[0022] One group of the slide rails and sliders that cooperate with each other is arranged between the support base and the adapter seat, and the other group of the slide rails and sliders is arranged between the mounting plate and the adapter seat.

[0023] In the technical solution of the embodiment of the present application, the movement of the mounting plate is achieved through multiple sets of sliding devices, thereby achieving precise control of the movement of the mounting plate.

[0024] In some embodiments, the verification piece includes:

[0025] a plate body having a calibration surface;

[0026] Multiple size verification parts are arranged on the verification surface at intervals along a straight line direction, and the multiple size verification parts have a length dimension in a first direction and a width dimension in a second direction. The length dimensions of the multiple size verification parts change in a gradient, and / or the width dimensions of the multiple size verification parts change in a gradient, wherein the first direction and the second direction are directions perpendicular to each other in a horizontal plane.

[0027] In the technical solution of the embodiment of the present application, multiple dimension calibration parts with gradient size changes are provided on the calibration piece, which effectively realizes the accuracy calibration of the measuring tools of 2D cameras and 3D cameras within the measuring range, ensuring that the linear offset of the measuring tools is calibrated. The linear offset is the linear error or endpoint linearity, which is the maximum deviation of the straight line composed of the endpoints of the entire measuring range, that is, the deviation between the measured curve and the ideal straight line. At the same time, it can calibrate the accuracy error caused by the distortion of the 2D camera lens and improve the accuracy of visual inspection.

[0028] In some embodiments, the size verification portion includes a protrusion or a groove formed on the plate body.

[0029] In the technical solution of the embodiment of the present application, the size information of the grooves, protrusions, etc. on the verification piece is obtained to determine whether it is within a reasonable error range with the predetermined size information, so as to achieve camera accuracy verification.

[0030] In some embodiments, the sizes of the plurality of size verification parts along the third direction vary in a gradient manner.

[0031] In the technical solution of the embodiment of the present application, a plurality of dimension calibration parts with gradient-changing dimensions are provided on the calibration piece, which effectively realizes the accuracy calibration within the measuring range of 2D camera and 3D camera measuring tools, and ensures that the linear offset of the measuring tool is calibrated.

[0032] In some embodiments, a color card calibration portion with a gradient grayscale value is further included on the calibration surface of the board body.

[0033] In the technical solution of the embodiment of the present application, the calibration of the imaging effect is achieved by providing a color comparison card on the calibration piece. The color card of the calibration piece requires that the black and white camera be equipped with three gradient standard color cards of light gray, gray, and dark gray, so as to achieve the calibration of the imaging effect of the black and white camera.

[0034] In some embodiments, the color card calibration unit includes a three-primary color card.

[0035] In the technical solution of the embodiment of the present application, the color camera is equipped with an RGB primary color card to achieve the calibration of the imaging effect of the color camera.

[0036] In some embodiments, the material of the plate body includes aluminum alloy; and / or,

[0037] The roughness of the calibration surface of the plate body is less than a preset roughness threshold.

[0038] In the technical solution of the embodiment of the present application, the material of the plate body includes aluminum alloy, and the roughness of the calibration surface of the plate body is adjusted to improve the imaging effect of multiple size calibration parts with gradient changes in size and the color card calibration part with gradient changes in grayscale values.

[0039] In a second aspect, the present application further provides a visual inspection method, wherein the visual inspection system includes a visual inspection device, a visual verification device, and a host computer, the visual verification device includes a verification piece and a verification piece moving device, the verification piece moving device includes a movably mounted mounting plate, the mounting plate can be in a verification position during the movable travel of the mounting plate, the verification position is within the detection range of the visual inspection device, and the verification piece is provided on the mounting plate;

[0040] The visual detection method comprises:

[0041] The visual inspection device obtains an image of the verification part and sends the image of the verification part to the host computer;

[0042] The host computer determines a systematic verification result of the visual inspection device based on the image of the verification part.

[0043] In the technical solution of the embodiment of the present application, when performing system calibration, the calibration parts are concentrated on the mounting plate, and the mounting plate is driven to move the calibration parts, thereby adopting an automatic method to achieve fully automatic calibration, thereby solving the problem that the calibration time is long and the production capacity loss is large due to the influence of the operator's proficiency.

[0044] In some embodiments, the verification component includes a plurality of size verification portions with gradient-varying sizes and a color card verification portion with gradient-varying grayscale values, and the visual inspection method further includes:

[0045] The visual inspection device obtains a first image of the size verification part and a second image of the color card verification part, and sends the first image and the second image to a host computer;

[0046] The host computer determines a systematic verification result of the visual inspection device based on the first image and / or the second image.

[0047] In the technical solution of the embodiment of the present application, a plurality of dimension calibration parts with gradient size changes are provided on the calibration piece, which effectively realizes the accuracy calibration within the measuring range of 2D cameras and 3D camera measuring tools, and ensures that the linear offset of the measuring tool is calibrated. The linear offset is the linear error or endpoint linearity, which is the maximum deviation of the straight line composed of the endpoints of the entire measuring range, that is, the deviation between the measured curve and the ideal straight line. At the same time, the accuracy error caused by the distortion of the 2D camera lens can be verified, and on the basis of achieving accuracy calibration, the algorithm calibration is further performed, thereby realizing the systematic calibration of the visual inspection device.

[0048] In some embodiments, the systematic verification results include visual accuracy verification results, imaging effect detection results, and visual algorithm verification results. The visual inspection method further includes:

[0049] In the event that any one of the visual accuracy verification result, the imaging effect detection result, and the visual algorithm verification result shows an abnormal result, the device to be detected is notified to stop working.

[0050] In the technical solution of the embodiment of the present application, in order to ensure the safety of production line production and prevent defective products from leaving the factory, when problems with visual accuracy, visual imaging and visual algorithms are detected on the production line, production is stopped for inspection in a timely manner to achieve effective management and control of production line monitoring.

[0051] In some embodiments, the visual detection method further comprises:

[0052] When the systematic verification result is a normal result, it is determined that the verification mode is ended, and the relevant production equipment is notified to enter the production mode.

[0053] In the technical solution of the embodiment of the present application, in order to avoid the presence of other materials in the verification and other influencing factors in the production process during the production verification, it is necessary to suspend the material delivery of the production line and enter the verification mode. However, after the verification is completed, production needs to be resumed as soon as possible and automatically restored to the production mode, thereby improving the automated management of the production line.

[0054] In some embodiments, the visual detection method further comprises:

[0055] When the host computer receives the verification instruction, it sends the verification instruction to the visual detection device;

[0056] When receiving the verification instruction, the visual inspection device notifies the station controller to stop conveying the material to the visual inspection device.

[0057] In the technical solution of the embodiment of the present application, in order to avoid the presence of other materials in the verification and other influencing factors in the production process during the production verification, the material delivery to the production line is suspended, thereby improving the accuracy of the verification.

[0058] In some embodiments, the visual detection method further comprises:

[0059] The visual inspection device notifies the driving device in the visual inspection device to move the verification piece to the verification position when detecting that no material is present at the verification position detected by the visual inspection device, and the verification position is within the detection range of the visual inspection device;

[0060] In the technical solution of the embodiment of the present application, the calibration piece is moved to the position to be measured by a driving device without manual operation of the calibration piece, thereby realizing fully automatic calibration and solving the problem that the calibration time is long and the production capacity is greatly lost due to the influence of the operator's proficiency.

[0061] In some embodiments, the visual detection method further comprises:

[0062] When the visual inspection device detects that there is material on the check position detected by the visual inspection device, it notifies the visual inspection device to continue to inspect the current material on the check position until the target material leaves the check position, and then notifies the driving device in the visual inspection device to move the check piece to the check position so that the light source in the visual inspection device illuminates the check piece.

[0063] In the technical solution of the embodiment of the present application, the visual inspection device uses a sensor to determine whether there is material at the position to be tested. If there is material, the last product inspection is performed, thereby realizing fully automatic verification and improving the efficiency of verification.

[0064] In some embodiments, the visual detection method further comprises:

[0065] When the first image and / or the second image is detected, the visual detection device notifies the driving device in the visual verification device to move the verification piece back to its original position.

[0066] In the technical solution of the embodiment of the present application, after obtaining the verification image, the verification piece is automatically moved back to its original position by the driving device without the need for manual operation of the verification piece, thereby realizing fully automatic verification and solving the problem of long verification time and large production capacity loss that is easily affected by the operator's proficiency.

[0067] In some embodiments, the visual detection method further comprises:

[0068] When the system check result is normal, the host computer notifies the station controller to start conveying materials;

[0069] When the system check result is abnormal, the host computer notifies each detection device to perform shutdown detection and issues an alarm.

[0070] In the technical solution of the embodiment of the present application, the upper computer controls the production line according to the systematic verification result. When the systematic verification result is normal, the workstation controller is notified to start conveying materials, thereby improving the processing efficiency of the production line. When the systematic verification result is abnormal, each detection equipment is notified to perform shutdown detection and issue an alarm, thereby improving the effective monitoring of the production line.

[0071] In some embodiments, the visual detection method further comprises:

[0072] The host computer performs visual accuracy verification based on the first image to obtain a visual accuracy verification result;

[0073] The host computer performs imaging effect verification based on the second image to obtain an imaging effect verification result;

[0074] When the visual accuracy verification result and the imaging effect verification result are both normal verification results, the host computer performs visual algorithm verification according to the target image to obtain a visual algorithm verification result;

[0075] The host computer determines a systematic verification result of the visual detection device according to the visual accuracy verification result, the imaging effect verification result and the visual algorithm verification result.

[0076] In the technical solution of the embodiment of the present application, by providing a plurality of dimension calibration parts with gradient-changing dimensions on the calibration piece, the accuracy calibration within the measuring range of the 2D camera and 3D camera measuring tools is effectively achieved, and on the basis of achieving the accuracy calibration, the algorithm calibration is further performed, thereby realizing the systematic calibration of the visual inspection device.

[0077] In some embodiments, the host computer performs visual accuracy verification based on the first image to obtain a visual accuracy verification result, including:

[0078] The visual accuracy verification result of the visual inspection device is determined according to the measurement values ​​corresponding to the structural parameters of the size verification part in the first image.

[0079] In the technical solution of the embodiment of the present application, the visual accuracy verification result is obtained by comparing the measurement values ​​corresponding to the structural parameters, thereby realizing the verification of visual accuracy through quantified data and improving the accuracy of visual detection.

[0080] In some embodiments, the host computer determines the visual accuracy verification result of the visual inspection device according to the measurement value corresponding to the structural parameter, including:

[0081] Determine a first difference between a measured value corresponding to the structural parameter and a preset standard value;

[0082] A visual accuracy verification result of the visual inspection device is determined according to the first difference.

[0083] In the technical solution of the embodiment of the present application, the visual accuracy verification result is obtained by the difference between the measured value corresponding to the structural parameter and the standard value corresponding to the actual detection object, thereby effectively obtaining the error between the image analysis result and the actual result, and more accurately obtaining the accuracy of visual detection.

[0084] In some embodiments, the host computer determines the visual accuracy verification result of the visual inspection device according to the first difference, including:

[0085] A visual accuracy verification result of the visual inspection device is determined according to the first difference and a first parameter threshold range.

[0086] In the technical solution of the embodiment of the present application, the visual accuracy verification result is obtained by comparing the difference between the measured value corresponding to the structural parameter and the standard value corresponding to the actual detection object with the parameter threshold range. Compared with comparison only by difference, by giving a certain threshold range, the detection result is considered abnormal only when it exceeds this range, thereby improving the accuracy of visual accuracy detection.

[0087] In some embodiments, the host computer performs visual accuracy verification based on the first image to obtain a visual accuracy verification result, including:

[0088] The host computer determines a first difference between a measurement value corresponding to a structural parameter of the size verification part in the first image and a preset standard value;

[0089] When the first difference is greater than or equal to the first parameter threshold range, the host computer obtains a verification result that the visual accuracy of the visual detection device is abnormal; when the first difference is less than the first parameter threshold range, the host computer obtains a verification result that the visual accuracy of the visual detection device is normal.

[0090] In the technical solution of the embodiment of the present application, the difference between the measured value corresponding to the structural parameter and the standard value corresponding to the actual detection object is compared with the parameter threshold range. The detection result is considered normal only if it does not exceed the range, thereby improving the accuracy of visual precision detection.

[0091] In some embodiments, it further includes:

[0092] The host computer determines a first parameter threshold range according to a tolerance corresponding to the structural parameter.

[0093] In the technical solution of the embodiment of the present application, before obtaining the parameter threshold range, the parameter threshold range is determined by corresponding structural parameters. Since different structural parameters correspond to different standards, the parameter threshold range is determined by the tolerance corresponding to the parameters, so that the parameter threshold range is adapted to the structural parameters, thereby improving the rationality of visual accuracy detection.

[0094] In some embodiments, the visual detection method further comprises:

[0095] The host computer determines a second difference between a measurement value corresponding to a structural parameter of the size verification part in the first image and a target standard value;

[0096] A first visual algorithm verification result of the visual inspection device is determined according to the second difference.

[0097] In the technical solution of the embodiment of the present application, the measurement values ​​corresponding to the structural parameters are compared with the measurement values ​​of the structural parameters analyzed by the normal visual detection algorithm to obtain the difference between the analysis results of the current visual algorithm and the analysis results detected by the normal visual detection algorithm. Therefore, the difference between the current visual algorithm and the normal visual detection algorithm is obtained, thereby realizing the verification of the visual algorithm.

[0098] In some embodiments, before determining the second difference between the measured value corresponding to the structural parameter and the standard value corresponding to the target structural parameter, the method further includes:

[0099] The first image is detected by a target vision algorithm to obtain target standard values ​​corresponding to the structural parameters.

[0100] In the technical solution of the embodiment of the present application, in order to more effectively obtain the difference between the current visual algorithm and the normal visual detection algorithm, the dimension verification part is analyzed by the target visual algorithm to obtain the target standard value corresponding to the structural parameter, which can be compared with the measurement value corresponding to the structural parameter, thereby realizing the standard unification of the comparison between the current visual algorithm and the normal visual detection algorithm, and improving the accuracy of the visual algorithm verification.

[0101] In some embodiments, determining a first visual algorithm verification result of the visual inspection device according to the second difference includes:

[0102] A first visual algorithm verification result of the visual detection device is determined according to the second difference and a second parameter threshold range.

[0103] In the technical solution of the embodiment of the present application, the visual algorithm verification result is obtained by comparing the difference between the measured value corresponding to the structural parameter and the standard value corresponding to the normal visual detection algorithm with the parameter threshold range. Compared with comparison only by difference, by giving a certain threshold range, the visual algorithm is considered abnormal only when it exceeds this range, thereby improving the accuracy of the visual algorithm verification.

[0104] In some embodiments, the visual detection method further comprises:

[0105] The host computer detects the first image using a target vision algorithm to obtain target standard values ​​corresponding to the structural parameters;

[0106] The host computer determines a second difference between a measurement value corresponding to a structural parameter of the size verification part in the first image and a target standard value;

[0107] When the second difference is greater than or equal to the second parameter threshold range, the host computer obtains a verification result that the visual algorithm of the visual detection device is abnormal; when the second difference is less than the second parameter threshold range, the host computer obtains a verification result that the visual algorithm of the visual detection device is normal.

[0108] In the technical solution of the embodiment of the present application, the visual algorithm verification result is obtained by comparing the difference between the measurement value corresponding to the structural parameter and the standard value corresponding to the normal visual detection algorithm with the parameter threshold range, thereby improving the accuracy of the visual algorithm verification.

[0109] In some embodiments, performing imaging effect verification according to the second image to obtain an imaging effect verification result includes:

[0110] An imaging effect verification result is obtained according to the grayscale value of the color card verification portion in the second image.

[0111] In the technical solution of the embodiment of the present application, by providing a colorimetric card on the calibration piece, the calibration of the imaging effect is further achieved on the basis of achieving visual accuracy.

[0112] In some embodiments, obtaining an imaging effect verification result according to the grayscale value of the color card verification portion in the second image includes:

[0113] determining a third difference between the grayscale value and a preset grayscale value;

[0114] A detection result of the imaging effect of the visual detection device is determined according to the third difference.

[0115] In the technical solution of the embodiment of the present application, the imaging effect verification result is obtained by the difference between the grayscale value and the grayscale value corresponding to the actual detection object, thereby effectively obtaining the error between the image analysis result and the actual result, and more accurately realizing the detection of the imaging effect.

[0116] In some embodiments, determining the detection result of the imaging effect of the visual detection device according to the third difference includes:

[0117] A detection result of the imaging effect of the visual detection device is determined according to the third difference and the third parameter threshold range.

[0118] In the technical solution of the embodiment of the present application, the imaging effect verification result is obtained by comparing the difference between the current grayscale value and the standard value corresponding to the actual detection object with the parameter threshold range. Compared with comparison only by difference, by giving a certain threshold range, the detection result is considered abnormal only when it exceeds this range, thereby improving the accuracy of imaging effect detection.

[0119] In some embodiments, the host computer performs imaging effect verification based on the second image to obtain an imaging effect verification result, including:

[0120] The host computer determines a third difference between the grayscale value of the color card verification part in the second image and a preset grayscale value;

[0121] When the third difference is greater than or equal to the third parameter threshold range, the host computer determines that the imaging effect of the visual detection device is abnormal; when the third difference is less than the third parameter threshold range, the host computer determines that the imaging effect of the visual detection device is normal.

[0122] In the technical solution of the embodiment of the present application, the imaging effect verification result is obtained by comparing the difference between the current grayscale value and the standard value corresponding to the actual detection object with the parameter threshold range. By giving a certain threshold range, the detection result is considered abnormal only when it exceeds this range, thereby improving the accuracy of imaging effect detection.

[0123] In some embodiments, the visual detection method further comprises:

[0124] A second visual algorithm verification result of the visual inspection device is obtained according to the grayscale value of the color card verification part in the second image.

[0125] In the technical solution of the embodiment of the present application, in addition to detecting the imaging effect through grayscale values, the visual algorithm can also be verified, thereby realizing systematic visual verification and improving the comprehensiveness and effectiveness of production line monitoring.

[0126] In some embodiments, obtaining a second visual algorithm verification result of the visual inspection device according to the grayscale value of the color card verification portion in the second image includes:

[0127] A second visual algorithm verification result of the visual inspection device is obtained according to a fourth difference between the grayscale value of the color card verification part in the second image and the target grayscale value.

[0128] In the technical solution of the embodiment of the present application, the grayscale value is compared with the grayscale value analyzed by the normal visual detection algorithm to obtain the difference between the analysis result of the current visual algorithm and the analysis result detected by the normal visual detection algorithm. Therefore, the difference between the current visual algorithm and the normal visual detection algorithm is obtained, thereby realizing the verification of the visual algorithm.

[0129] In some embodiments, before obtaining the second visual algorithm verification result of the visual inspection device based on the fourth difference between the grayscale value of the color card verification portion in the second image and the target grayscale value, the method further includes:

[0130] The second image is detected by a target vision algorithm to obtain a target grayscale value.

[0131] In the technical solution of the embodiment of the present application, in order to more effectively obtain the difference between the current visual algorithm and the normal visual detection algorithm, the colorimetric card is analyzed by the target visual algorithm to obtain the target standard value corresponding to the colorimetric card, which can be compared with the grayscale value obtained by the current visual algorithm analysis, thereby achieving the standard unification for the comparison between the current visual algorithm and the normal visual detection algorithm, and improving the accuracy of the visual algorithm verification.

[0132] In some embodiments, obtaining a second visual algorithm verification result of the visual inspection device according to a fourth difference between the grayscale value of the color card verification portion in the second image and the target grayscale value includes:

[0133] A second visual algorithm verification result of the visual detection device is determined according to the fourth difference and the fourth parameter threshold range.

[0134] In the technical solution of the embodiment of the present application, the visual algorithm verification result is obtained by comparing the difference between the grayscale value corresponding to the current visual algorithm and the standard value corresponding to the normal visual detection algorithm with the parameter threshold range. Compared with comparison only by difference, by giving a certain threshold range, the visual algorithm is considered abnormal only when it exceeds this range, thereby improving the accuracy of the visual algorithm verification.

[0135] In some embodiments, the visual detection method further comprises:

[0136] The host computer detects the second image using a target vision algorithm to obtain a target grayscale value;

[0137] The host computer determines a fourth difference between the grayscale value of the color card verification part in the second image and the target grayscale value;

[0138] The host computer obtains a verification result of an abnormality in the visual algorithm of the visual detection device when the fourth difference is greater than or equal to a fourth parameter threshold range;

[0139] When the fourth difference is less than the fourth parameter threshold range, the host computer obtains a verification result that the vision algorithm of the vision detection device is normal.

[0140] In the technical solution of the embodiment of the present application, the visual algorithm verification result is obtained by comparing the difference between the grayscale value corresponding to the current visual algorithm and the standard value corresponding to the normal visual detection algorithm with the parameter threshold range. Compared with comparison only by difference, by giving a certain threshold range, the visual algorithm is considered abnormal only when it exceeds this range, thereby improving the accuracy of the visual algorithm verification.

[0141] In some embodiments, the visual detection method further comprises:

[0142] A systematic verification result of the visual inspection device is determined based on the detection value and the standard value corresponding to the first image.

[0143] In the technical solution of the embodiment of the present application, the systematic verification result of the visual inspection device is obtained by comparing the detection value with the standard value, thereby realizing the systematic verification result of the visual inspection device through quantified data and improving the accuracy of the systematic verification of the visual inspection device.

[0144] In some embodiments, the visual detection method further comprises:

[0145] A systematic verification result of the visual inspection device is determined based on the detection value and the standard value corresponding to the second image.

[0146] In the technical solution of the embodiment of the present application, the visual accuracy verification result is obtained by comparing the measured value corresponding to the structural parameter with the standard value, and the imaging effect verification result is obtained by comparing the grayscale value of the colorimetric card with the standard value, thereby realizing the systematic verification of the visual detection device through quantified data and improving the accuracy of the systematic verification.

[0147] In some embodiments, the visual detection method further comprises:

[0148] A systematic verification result of the visual inspection device is determined based on the detection value and the standard value corresponding to the first image and the detection value and the standard value corresponding to the second image.

[0149] In the technical solution of the embodiment of the present application, visual accuracy verification and imaging effect verification are achieved simultaneously through the images of multiple size verification parts with gradient changes in size and color card verification parts with gradient changes in grayscale values. Then, the visual algorithm is verified based on the visual accuracy verification and imaging effect verification, thereby realizing systematic verification of the visual detection device.

[0150] In some embodiments, the host computer performs visual algorithm verification based on the target image, and before obtaining the visual algorithm verification result, further includes:

[0151] The host computer obtains a first picture, a second picture, a third picture, and a fourth picture with calibrated parameter values, wherein the parameter value of the first picture is within a first range, the parameter value of the second picture is within a second range, the parameter value of the third picture is within a third range, and the parameter value of the third picture is within a fourth range, the first range and the second range are different, and the third range and the fourth range are different;

[0152] A picture verification library is established based on the first picture, the second picture, the third picture and the fourth picture.

[0153] In the technical solution of the embodiment of the present application, the visual algorithm is detected by using a pre-established image verification library, which is more efficient than directly comparing the verification results.

[0154] In some embodiments, the host computer establishes an image verification library based on the first image, the second image, the third image, and the fourth image, including:

[0155] The host computer numbers the first picture, the second picture, the third picture, and the fourth picture;

[0156] An image verification library is established based on the numbered first image, second image, third image, and fourth image.

[0157] In the technical solution of the embodiment of the present application, when establishing the image verification library, the image samples are managed by numbering, thereby achieving effective management of the image verification library, and it is not convenient for subsequent adjustment and update of the image verification library.

[0158] In some embodiments, before obtaining the first picture, the second picture, the third picture, and the fourth picture with calibrated parameter values, the method further includes:

[0159] The sample images are evaluated using a target visual detection algorithm to obtain first, second, third, and fourth images with calibrated parameter values. In the technical solution of the embodiment of the present application, the sampled images are pre-evaluated using a complete visual detection algorithm, and the evaluated parameter values ​​are used as standard values ​​for subsequent comparisons to achieve visual algorithm verification.

[0160] In some embodiments, the host computer performs visual algorithm verification based on the target image to obtain a visual algorithm verification result, including:

[0161] The host computer selects a target image from the image verification library;

[0162] Performing image recognition on the target image by the visual detection device to obtain parameter values ​​of the target image;

[0163] comparing the parameter value with a calibration parameter value corresponding to the target image;

[0164] When the difference between the parameter value and the calibration parameter value corresponding to the target image exceeds the fifth parameter threshold range, it is determined that the visual algorithm is abnormal.

[0165] In the technical solution of the embodiment of the present application, the parameter values ​​detected by the improved visual detection algorithm are compared with the parameters detected by the current visual algorithm, and the visual algorithm is verified based on the comparison results to improve the accuracy of the visual algorithm verification. Beneficial effects

[0166] In the technical solution of the embodiment of the present application, when performing system calibration, the calibration parts are concentrated on the mounting plate, and the mounting plate is driven to move the calibration parts, thereby adopting an automatic method to achieve fully automatic calibration, thereby solving the problem that the calibration time is long and the production capacity loss is large due to the influence of the operator's proficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0167] FIG1 is a schematic diagram of the structure of a visual verification system proposed in some embodiments of the present application;

[0168] FIG2 is a schematic diagram of the structure of a verification component proposed in some embodiments of the present application;

[0169] FIG3 is a schematic diagram of the structure of a visual verification device proposed in some embodiments of the present application;

[0170] FIG4 is another schematic diagram of the structure of a verification piece proposed in some embodiments of the present application;

[0171] FIG5 is a schematic diagram of a first flow chart of a visual inspection method proposed in some embodiments of the present application;

[0172] FIG6 is a schematic diagram of a second flow chart of the visual inspection method proposed in some embodiments of the present application;

[0173] FIG7 is a schematic diagram of a third flow chart of the visual inspection method proposed in some embodiments of the present application;

[0174] FIG8 is a schematic diagram of a first overall process of visual inspection proposed in some embodiments of the present application;

[0175] FIG9 is a schematic diagram of a fourth flow chart of the visual inspection method proposed in some embodiments of the present application;

[0176] FIG10 is a schematic diagram of a fifth flow chart of the visual inspection method proposed in some embodiments of the present application;

[0177] FIG11 is a sixth flow chart of the visual inspection method proposed in some embodiments of the present application;

[0178] FIG12 is a seventh flow chart of the visual inspection method proposed in some embodiments of the present application;

[0179] FIG13 is a schematic diagram of an eighth flow chart of the visual inspection method proposed in some embodiments of the present application;

[0180] FIG14 is a ninth flow chart of the visual inspection method according to some embodiments of the present application;

[0181] FIG15 is a schematic diagram of a second overall flow chart of the visual inspection method proposed in some embodiments of the present application;

[0182] FIG16 is a schematic diagram of a visual algorithm detection process of a visual detection method proposed in some embodiments of the present application;

[0183] FIG17 is a schematic diagram of the tenth flow chart of the visual inspection method proposed in some embodiments of the present application;

[0184] FIG18 is a schematic diagram of an eleventh flow chart of a visual inspection method proposed in some embodiments of the present application;

[0185] FIG19 is a twelfth flow chart of the visual inspection method proposed in some embodiments of the present application;

[0186] FIG20 is a schematic diagram of a thirteenth flow chart of a visual inspection method proposed in some embodiments of the present application;

[0187] FIG21 is a third overall flow chart of the visual inspection method proposed in some embodiments of the present application.

[0188] The figure numbers in the specific implementation method are as follows: visual verification device 100, visual inspection device 200, host computer 300, verification piece 10, verification piece moving device 20, size verification part 201, color card verification part 202, plate body 30, verification surface 40, mounting plate 50, support base 60, driving device 70, 701, 702, slide rail 80, adapter 90, height limit block 901, first sliding limit block 902 and second sliding limit block 903. Modes for Carrying Out the Invention

[0189] The following embodiments of the technical solution of the present application will be described in detail with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present application and are therefore only examples and are not intended to limit the scope of protection of the present application.

[0190] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned figure descriptions are intended to cover non-exclusive inclusions.

[0191] In the description of the embodiments of this application, technical terms such as "first" and "second" are used solely to distinguish different objects and should not be understood to indicate or imply relative importance or to implicitly specify the quantity, specific order, or primary and secondary relationship of the technical features indicated. In the description of the embodiments of this application, "plurality" means more than two, unless otherwise specifically defined.

[0192] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0193] In the description of the embodiments of this application, the term "and / or" is simply a description of the association relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent the following three situations: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this document generally indicates that the associated objects are in an "or" relationship.

[0194] On chip manufacturing lines, equipment inspection is a process used to check and evaluate the status of manufacturing equipment to ensure proper functioning and product quality. Equipment inspection on chip manufacturing lines typically requires greater rigor and precision, as chip manufacturing places extremely high demands on equipment precision, stability, and purity.

[0195] The following are the general steps for equipment inspection on the chip manufacturing production line: Cleaning and preparation: Before conducting the inspection, the equipment must first be cleaned and the work area must be kept tidy and clean. In addition, prepare the required inspection tools, record forms, and documents. Appearance inspection: Carefully inspect the appearance of the equipment, including checking whether there is damage, corrosion, and dirt on the surface of the equipment. Ensure that the mechanical parts and interfaces of the equipment are firmly connected to prevent loosening or leakage. Functional testing: Test the function of the equipment, including starting, running, and stopping the equipment. Check whether the equipment is working properly according to the predetermined parameters and procedures. Ensure that all functions can be performed accurately. Temperature and humidity control: Precise control of temperature and humidity is very important in the chip manufacturing process. Ensure that the temperature and humidity control systems of the equipment are operating normally and meet manufacturing requirements.

[0196] Sensor Inspection and Calibration: Chip manufacturing equipment is typically equipped with numerous sensors that monitor various parameters, such as temperature, pressure, and flow. Check the accuracy of these sensors and perform calibration and adjustments. Check and Replace Consumables: Chip manufacturing equipment typically uses a variety of consumables, such as filters and seals. Check the condition of these consumables and, if necessary, replace them promptly to ensure proper equipment operation and maintain the purity of the manufacturing process. Recording and Reporting: During spot checks, record inspection results, observed issues, and actions taken. Report any serious issues promptly and take appropriate corrective action.

[0197] Equipment inspection is crucial in chip manufacturing production lines, helping to ensure accuracy, stability, and reliability, thereby guaranteeing high quality during the chip manufacturing process and ensuring product performance. Regular equipment inspection is also a key task in managing and maintaining chip manufacturing equipment. Typically, this can be performed using a vision system combined with inspection modules.

[0198] On chip manufacturing production lines, the timing of equipment inspections is typically based on the following factors: Regular inspections: Regular inspections are performed at predetermined intervals, such as daily, weekly, or monthly. These inspections ensure equipment stability and reliability and detect potential problems early. Preventive inspections: Preventive inspections are performed during normal equipment operation to prevent problems that could lead to failure or damage. These inspections can reduce equipment failures and downtime, improving production efficiency. Occasional inspections: Occasional inspections are performed when abnormal conditions occur or equipment problems arise. These inspections aim to promptly identify and resolve specific issues and prevent them from escalating. Critical node inspections: Critical node inspections are performed before or after critical nodes or important steps in the chip manufacturing process. This ensures that the equipment is in good condition at critical moments and avoids negative impacts on product quality. Furthermore, the timing of equipment inspections can be determined based on the equipment's historical record, manufacturing needs, and operational experience. It is important to implement a reasonable inspection plan to ensure the stability and reliability of the equipment at critical moments and critical steps, and to minimize risks and failures in production. Generally, the manufacturing execution system creates a verification task to trigger the vision system to enter the automatic verification mode. The automatic verification time can also be set by the equipment to enter the verification mode.

[0199] For production line calibration, traditional measurement tool calibration is performed manually. In addition to being cumbersome and requiring skilled workers, the operation and analysis are time-consuming, resulting in significant production capacity losses for wire drawing.

[0200] In order to solve the technical problem that the existing verification method is time-consuming, the embodiment of the present application uses a visual verification device and a verification piece moving device installed for visual verification to achieve automated verification, thereby improving the efficiency of verification.

[0201] This application addresses the technical problem that existing verification methods are time-consuming. A visual inspection system is proposed. As shown in FIG1 , the visual inspection system includes:

[0202] The visual inspection device 100 includes a verification object 10 and a verification object moving device 20. The verification object moving device 20 includes a movably mounted mounting plate 50. During its travel, the mounting plate 50 can be positioned in a verification position, which is within the detection range of the visual inspection device 200. The verification object 10 is mounted on the mounting plate 50. The visual inspection device 200 is configured to obtain an image of the verification object 10 and transmit the image of the verification object 10 to a host computer 300.

[0203] The host computer 300 is used to determine the systematic verification result of the visual inspection device 200 based on the image of the verification object 10.

[0204] In this embodiment, during the battery manufacturing process, the quality monitoring of battery cells, modules, and PACKs is becoming increasingly sophisticated. Therefore, a visual inspection system is provided in each process to perform quality monitoring through the visual inspection system. This embodiment can be applied to each process link of battery cells, modules, and PACKs. As long as a visual inspection system is involved, this embodiment can also achieve automated inspection. Taking the battery cell welding process as an example, after the battery cell completes the welding process, the welded battery cell is transported to the next process. A visual inspection device 200 is arranged on the transportation line for transporting the welded battery cell. The visual inspection device 200 is used to visually monitor the production line. Therefore, there is a check position for the visual inspection device 200 during transportation. When the welded battery cell reaches the check position during transportation, the visual inspection device 200 can capture the image of the welded battery cell and perform data analysis based on the image of the welded battery cell. In this embodiment, when the visual inspection system is in inspection mode, the check piece 10 on the visual inspection device 100 is moved to the check position, and the visual inspection of the check position is performed by the visual inspection system, thereby realizing automated inspection and achieving the purpose of improving the efficiency of production line inspection.

[0205] The visual verification device 100 can be located on one side of the verification position. When the manufacturing execution system creates a verification task to trigger the visual system to enter the automatic verification mode or the detection equipment sets the automatic verification time to enter the verification mode, the host computer 300 notifies the visual verification device 100 to enter the verification mode. The visual verification device 100 moves the verification piece 10 to the verification position through the verification piece moving device 20 for visual inspection. Since the visual inspection device 200 is used to collect the image of the inspection object located at the verification position, the visual inspection device 200 can collect the image of the verification piece 10 and send the collected verification piece 10 to the host computer 300 for image analysis. The host computer 300 performs a systematic verification based on the image of the verification piece 10 to obtain a systematic inspection result, wherein the systematic inspection result includes the inspection result of whether the visual accuracy is abnormal, the inspection result of whether the imaging effect is abnormal, and the inspection result of whether the visual algorithm is abnormal.

[0206] The visual inspection system of this embodiment is adaptable to different application scenarios of visual inspection equipment. The device itself can have an automatic calibration function according to the characteristics of the measured object, and can automatically realize online calibration of the visual system light source, camera accuracy, and algorithm stability accuracy.

[0207] In the technical solution of the embodiment of the present application, the verification piece 10 is concentrated on the mounting plate 50, and the movement of the verification piece 10 is achieved by driving the mounting plate 50, thereby adopting an automatic method to achieve fully automatic verification, thereby solving the problem that the verification time is long and the production capacity loss is large due to the influence of the operator's proficiency.

[0208] In this embodiment, the verification piece moving device 20 is used to move the verification block to the verification position so that the light source of the visual inspection device 200 illuminates the verification piece 10. The verification piece moving device 20 is used to move the verification piece 10 under the light source for visual inspection when verification is required, thereby avoiding affecting the normal operation of the production line. When verification is not required, the verification piece 10 is automatically retracted. As shown in the structural diagram of the visual inspection device 100 in Figure 1, when verification is required, the verification piece 10 on the verification piece moving device 20 is moved under the light source for visual inspection, thereby achieving automated visual inspection.

[0209] The check digit is the position of visual inspection. The visual inspection device 200 includes a light source, a visual inspection component, a processing device and a control device. The visual inspection component faces the check digit and is used to collect the image of the product on the check digit. The light source provides light for the check digit. The processing device is used to obtain the visual image. The control device is used to realize the workflow control of the visual inspection device 200 on the production line. By arranging visual inspection on the check digit, effective management and control of the production line can be achieved.

[0210] In this embodiment, when system calibration is required using the calibration piece 10, the calibration piece 10 is automatically moved to the calibration position, thereby achieving fully automatic calibration of the visual inspection device 200. This embodiment centralizes the calibration piece 10 on the mounting plate 50 and drives the mounting plate 50 to move the calibration piece 10, thereby achieving fully automatic calibration in an automated manner. This solves the problem of long calibration times and significant production capacity losses that are easily affected by operator proficiency.

[0211] According to some embodiments of the present application, optionally, the verification piece 10 can be detachably mounted on the mounting plate 50 .

[0212] In this embodiment, the verification piece 10 and the mounting plate 50 can be connected by a lock or other methods, which are not limited in this embodiment. The verification piece 10 can be detachably mounted on the mounting plate 50. Therefore, the verification piece 10 on the mounting plate 50 can be replaced according to actual detection needs, thereby improving the flexibility of detection.

[0213] In this embodiment, the quick-release locking method of the calibration member 10 is achieved by using the detachable parts on the mounting plate 50 of the calibration member 10, which is convenient for true value measurement of a two-dimensional or higher precision measurement system.

[0214] According to some embodiments of the present application, optionally, the calibration member 10 includes a plurality of size calibration portions with gradient-varying sizes and a color card calibration portion with gradient-varying grayscale values;

[0215] The visual inspection device 200 is further configured to obtain a first image of the size verification portion and a second image of the color card verification portion, and send the first image and the second image to the host computer 300;

[0216] The host computer 300 is further configured to determine a systematic verification result of the visual inspection device 200 based on the first image and / or the second image.

[0217] As shown in Figure 2, the calibration piece 10 and the calibration part of the calibration piece 10 are schematically shown. The calibration part includes multiple size calibration parts 201 with gradient changes in size and a color card calibration part 202 with gradient changes in grayscale value. The size calibration part can be a gradient calibration groove or protrusion, or a size calibration part in other forms. The calibration piece 10 requires that the size of the groove or protrusion presents a periodic regularity and covers the measuring range as much as possible. The calibration piece 10 requires that the groove or protrusion is not a mirror surface, but a frosted surface with a roughness of <Ra3.2。

[0218] Under normal circumstances, the dimensional information of the grooves, protrusions, etc. on the calibration piece 10 only has a gradient change in the depth dimension, while the width or length has no gradient change. There is a certain degree of randomness in the calibration of the 2D camera, and it is impossible to effectively calibrate and confirm the linear offset of the measuring range, and it is impossible to detect the accuracy impact caused by lens distortion. Therefore, this embodiment adds gradient changes in two-dimensional and three-dimensional dimensions to effectively realize the accuracy calibration within the measuring range of 2D cameras and 3D cameras, ensure that the linear offset of the measuring tool is calibrated, and at the same time, it can calibrate the accuracy error caused by the distortion of the 2D camera lens.

[0219] In this embodiment, the calibration piece 10 requires no less than 5 grooves or protrusions with gradient changes, and at the same time satisfies 2D and 3D dimensional gradient changes. It is suitable for the accuracy calibration of the measuring range of 2D cameras and 3D camera measuring tools, and is suitable for linear offset analysis and calibration of 2D cameras and 3D camera measuring tools.

[0220] In a specific implementation, the first image is a captured image of multiple size verification parts with gradient changes in size. The verification piece 10 can also be placed on a product to obtain an image of the product, and the visual algorithm is verified based on the image of the product. A visual inspection device 200 is placed directly opposite the verification piece 10. When the verification piece 10 is supplemented with light by a light source, the visual inspection device 200 obtains an image of the size verification part on the verification piece 10. A systematic analysis is performed based on the image of the size verification part to achieve detection by the visual inspection device 200. The second image is a captured image of a color card verification part with gradient changes in grayscale values. The first image and the second image can be a single image or multiple separate images. When there are multiple images, the first and second images of the corresponding areas are captured by visual inspection devices 200 arranged in different areas. When there is a single image, the image can be divided into regions to obtain the first and second images corresponding to the multiple size verification parts with gradient changes in size and the color card verification part with gradient changes in grayscale values, thereby achieving visual inspection.

[0221] The verification piece 10 can also be placed on a product to obtain an image of the product, and visual algorithms can be verified based on the product image. A visual inspection device 200 is positioned directly opposite the verification piece 10. By supplementing the verification piece 10 with light, the visual inspection device 200 captures an image of the dimensional verification portion of the verification piece 10. This image of the dimensional verification portion is systematically analyzed to enable inspection by the visual inspection device 200.

[0222] In this embodiment, the systematic verification result includes a visual accuracy verification result, an imaging effect verification result, and a visual algorithm verification result, thereby achieving a systematic verification of the visual inspection device 200.

[0223] In this embodiment, the calibration component 10 is provided with multiple dimensional calibration sections with gradient variations in size, effectively achieving accuracy calibration within the measuring range of 2D and 3D camera measuring tools, ensuring that the linear offset of the measuring tools is verified. Linear offset is a linear error or endpoint linearity. It is the maximum deviation of the line formed by the endpoints of the entire measuring range, that is, the deviation between the measured curve and the ideal line. It can also verify the accuracy error caused by distortion of the 2D camera lens. Based on the accuracy verification, further algorithm verification is performed, thereby achieving systematic verification of the visual inspection device 200.

[0224] According to some embodiments of the present application, optionally, as shown in FIG3 , the visual verification device 100 further includes:

[0225] a support base 60; and

[0226] The driving device 70 includes a fixed portion 701 and a movable portion 702 that move linearly relative to each other. The fixed portion 701 is disposed on the support base 60 , and the movable portion 702 is connected to the mounting plate 50 .

[0227] The fixing portion 60 is used to fix the driving device on the support base 60. The fixing portion 60 can be a fixing part or other forms of fixing parts, and this embodiment does not limit this. The driving device 70 can be a cylinder or other driving devices, and this embodiment does not limit this. The movable portion 702 can be a telescopic rod of the cylinder, one end of the telescopic rod is connected to the fixing part, and the other end is connected to the mounting plate 50. The movement of the mounting plate 50 is achieved by the telescopic rod, thereby driving the verification piece 10 on the mounting plate 50 to move to the verification position.

[0228] In this embodiment, the driving device 70 does not directly drive the verification piece 10, but drives the mounting plate 50 connected to the verification piece 10. By driving the mounting plate 50, the movement of the verification piece 10 is achieved. The mounting plate 50 is connected to the verification piece 10 through a detachable part, so that the verification piece 10 can be replaced according to needs, thereby improving the flexibility of visual inspection.

[0229] In a specific implementation, the driving device 70 is taken as an example of a cylinder. One end of the cylinder is connected to the mounting plate 50, and the verification piece 10 is installed on the mounting plate 50 through a detachable part. When the cylinder receives a detection instruction, it is actuated to move the connected mounting plate 50 to the verification position, thereby causing the verification piece 10 to automatically move to the verification position. When the cylinder receives a detection completion instruction, it retracts and moves the connected verification piece 10 mounting plate 50 back to its original position, thereby causing the verification piece 10 to automatically move back to its original position.

[0230] The driving device 70 can also be directly connected to the verification piece 10. When the cylinder receives a detection instruction, it is actuated to move the connected verification piece 10 to the verification position, so that the verification piece 10 automatically moves to the verification position. When the cylinder receives a detection completion instruction, it retracts and moves the connected verification piece 10 back to its original position, so that the verification piece 10 automatically moves back to its original position.

[0231] In the technical solution of the embodiment of the present application, the movement of the verification piece 10 is achieved by driving, thereby realizing full-automatic verification of the visual inspection device 200.

[0232] According to some embodiments of the present application, optionally, the visual verification device 100 also includes a sliding guide structure, which includes a slide rail 80 and a slider (not shown in the figure) that slide with each other. Among the slide rails 80 and the slider, the slide rails 80 and the slider are arranged between the support base and the mounting plate 50, one of which is arranged on the support base 60 and the other is arranged on the mounting plate 50.

[0233] In this embodiment, a sliding guide structure is used in conjunction with a sliding device located at the bottom of the mounting plate 50; the sliding guide structure is used to move the verification piece 10 to the verification position. A slider is mounted on the support base 60, and a slide rail 80 moves on the slider. The slide rail 80 cooperates with the mounting plate 50. When the drive device 70 drives the mounting plate 50, the slide rail 80 moves on the slider, thereby moving the verification piece 10 on the mounting plate 50 to the verification position.

[0234] In this embodiment, when the mounting plate 50 of the verification piece 10 is driven by the driving device 70 , the movement trajectory of the mounting plate 50 of the verification piece 10 is determined by the sliding device, thereby improving the accuracy of the movement of the verification piece 10 .

[0235] According to some embodiments of the present application, optionally, the visual verification device 100 further includes an adapter 90 , two slide rails are provided, and two corresponding sliders are provided;

[0236] One set of slide rails 80 and sliders that cooperate with each other is disposed between the support base 60 and the adapter seat 90 , and another set of slide rails 80 and sliders is disposed between the mounting plate 50 and the adapter seat.

[0237] In this embodiment, as shown in Figure 3, a double track is provided, and a first set of mutually slidingly matched slide rails 80 and sliders are provided on the side of the driving device 70 close to the calibration block. The calibration block is controlled to move within a first moving range by the first set of slide rails 80 and sliders, and a second set of mutually slidingly matched slide rails 80 and sliders are provided on the side away from the calibration block. The calibration block is controlled to move within a second moving range by the second set of slide rails 80 and sliders. The calibration block is moved to the calibration position by the first set of mutually slidingly matched slide rails 80 and sliders, and the calibration block can be retracted to the support base 60 by the second set of mutually slidingly matched slide rails 80 and sliders.

[0238] In this embodiment, the movement of the mounting plate 50 is achieved through multiple sets of sliding devices, thereby achieving precise control of the movement of the mounting plate 50.

[0239] The adapter is also equipped with a height limiter 901 and a first sliding limiter 902. The height limiter 901 is used to limit the height of the sliding track, restricting the vertical sliding range of the mounting plate 50. The first sliding limiter 902 is used to limit the horizontal sliding range of the slide rail within the first range of movement. The visual inspection device 100 also includes a second sliding limiter 903, which is fixed to the support base 60 and is used to limit the horizontal sliding range of the slide rail within the second range of movement.

[0240] According to some embodiments of the present application, optionally, as shown in FIG4 , the visual inspection verification piece 10 includes:

[0241] The plate body 30 has a calibration surface 40;

[0242] Multiple size verification parts 201 are arranged on the verification surface 40 at intervals along a straight line direction. The multiple size verification parts have a length dimension in a first direction and a width dimension in a second direction. The length dimensions of the multiple size verification parts change in a gradient, and / or the width dimensions of the multiple size verification parts change in a gradient, wherein the first direction and the second direction are directions perpendicular to each other in a horizontal plane.

[0243] In this embodiment, the verification piece 10 includes a verification block, and may also be a verification piece 10 in other forms. This embodiment does not impose any restrictions on this. Taking the verification block as an example, the first direction may be the horizontal direction along the plate body 30, that is, the x-axis direction, and the second direction may be the vertical direction along the plate body 30, that is, the y-axis direction. It may also be that the first direction is along the vertical direction of the plate body 30, and the second direction is along the horizontal direction of the plate body 30. This embodiment does not impose any restrictions on this.

[0244] In a specific implementation, the length dimension of the dimension verification part on the front side of any two adjacent dimension verification parts is greater than or smaller than the length dimension of the dimension verification part on the rear side, and / or the width dimension of the dimension verification part on the front side of any two adjacent dimension verification parts is greater than or smaller than the width dimension of the dimension verification part on the rear side, wherein the first direction and the second direction are directions perpendicular to each other in the horizontal plane, so that the length or width dimension changes linearly, thereby realizing the verification of linear offset.

[0245] In the spacing direction of the multiple dimension verification sections, among any three adjacent dimension verification sections, the length change of the two adjacent dimension verification sections on the front side is the same as the length change of the two adjacent dimension verification sections on the rear side; and / or the width change of the two adjacent dimension verification sections on the front side is the same as the width change of the two adjacent dimension verification sections on the rear side. This results in a gradient change in the length or width, improving the accuracy of the linear offset.

[0246] In this embodiment, the size calibration portion and the plate body 30 may be integrally formed or may be separately provided, and this embodiment does not impose any limitation on this.

[0247] In this embodiment, a plurality of dimension calibration parts with gradient-changing dimensions are provided on the calibration piece 10, which effectively realizes the accuracy calibration within the measuring range of 2D camera and 3D camera measuring tools, and ensures that the linear offset of the measuring tools is calibrated. The linear offset is the linear error or endpoint linearity, which is the maximum deviation of the straight line composed of the endpoints of the entire measuring range, that is, the deviation between the measured curve and the ideal straight line. At the same time, it can calibrate the accuracy error caused by the distortion of the 2D camera lens, thereby improving the accuracy of visual inspection.

[0248] According to some embodiments of the present application, optionally, the size calibration portion includes a protrusion or a groove formed on the plate body.

[0249] In this embodiment, the size verification part is provided with a protrusion or a groove, thereby increasing the visual accuracy verification on the basis of two-dimensional and three-dimensional, and its size changes in a gradient, effectively realizing the accuracy verification within the measuring range of 2D cameras and 3D camera measuring tools, ensuring that the linear offset of the measuring tool is verified, and at the same time, the accuracy error caused by the distortion of the 2D camera lens can be verified.

[0250] This embodiment obtains whether the size information of the grooves, protrusions, etc. on the verification piece 10 is within a reasonable error range with the predetermined size information, thereby achieving camera accuracy verification.

[0251] According to some embodiments of the present application, optionally, the sizes of the multiple size verification parts along a third direction change in a gradient, and the third direction is perpendicular to the first direction and the second direction.

[0252] In this embodiment, the third direction is along the z-axis direction of the plate body 30, that is, the z-axis direction of the dimension calibration part changes in a gradient.

[0253] In this embodiment, a plurality of dimension calibration parts with dimensions changing in a gradient are provided on the calibration part 10, effectively realizing the accuracy calibration within the measuring ranges of 2D cameras and 3D camera measuring tools, and ensuring that the linear offset of the measuring tool is calibrated.

[0254] According to some embodiments of the present application, optionally, it further includes a color card calibration part 202 with a gradient change in gray value provided on the calibration surface 40 of the plate body 30.

[0255] In this embodiment, the color card calibration part 202 can be a color comparison card, or other forms of color cards, which are not limited in this embodiment. The color comparison card includes standard color cards with multiple gradients. The color comparison card can be standard color cards with three gradients of light gray, gray, and dark gray, or an RGB primary color card. The black and white camera is equipped with standard color cards with three gradients of light gray, gray, and dark gray, and the color camera is equipped with an RGB primary color card for calibrating the imaging effect.

[0256] In this embodiment, by providing a color comparison card on the calibration part 10, the imaging effect is calibrated. The color card of the calibration part 10 requires the black and white camera to be equipped with standard color cards with three gradients of light gray, gray, and dark gray, so as to calibrate the imaging effect of the black and white camera.

[0257] According to some embodiments of the present application, optionally, the color card calibration part 202 includes a primary color card.

[0258] In this embodiment, the color camera is equipped with an RGB primary color card, thereby realizing the calibration of the imaging effect of the color camera.

[0259] According to some embodiments of the present application, optionally, the material of the plate body 30 includes aluminum alloy; and / or, the roughness of the calibration surface 40 of the plate body 30 is less than a preset roughness threshold.

[0260] In this embodiment, the calibration block in the vision detection system includes an aluminum white base material, such as an aluminum alloy material, gradient calibration grooves or protrusions, gradient standard color cards, a cylinder required for the movement of the calibration block, a slide rail 80, and a slider. The preset roughness threshold can be Ra3.2, or other thresholds, which are not limited in this embodiment. The calibration block requires the sizes of the grooves or protrusions to show a periodic pattern, covering the measuring range of the measuring tool as much as possible, and the calibration block requires the grooves or protrusions not to be mirror-like, with a matte texture on the surface and a roughness <Ra3.2, so as to obtain a better imaging effect.

[0261] In this embodiment, the material of the plate body 30 includes aluminum alloy, and the roughness of the calibration surface 40 of the plate body 30 is adjusted to improve the imaging effect of the multiple size calibration parts with gradient changes in size and the color card calibration part 202 with gradient changes in grayscale values.

[0262] The present application also proposes a visual inspection method, wherein the visual inspection system includes a visual inspection device, a visual verification device, and a host computer. As shown in FIG1 , the visual verification device includes a verification piece and a verification piece moving device. The verification piece moving device includes a movably mounted mounting plate. During the movable travel of the mounting plate, the mounting plate can be in a verification position. The verification position is within the detection range of the visual inspection device. The verification piece is arranged on the mounting plate.

[0263] As shown in FIG5 , which is a flow chart of a first embodiment of a visual inspection method, the visual inspection method includes:

[0264] Step S10: The visual inspection device obtains an image of the verification part and sends the image of the verification part to the host computer;

[0265] In step S20 , the host computer determines the systematic verification result of the visual inspection device according to the image of the verification part.

[0266] It should be noted that the visual verification device can be located on one side of the verification position. When the manufacturing execution system creates a verification task to trigger the visual system to enter the automatic verification mode or the detection equipment sets the automatic verification time to enter the verification mode, the host computer notifies the visual verification device to enter the verification mode. The visual verification device moves the verification piece to the verification position through the verification piece moving device for visual inspection. Since the visual inspection device is used to collect the image of the inspection object located at the verification position, the visual inspection device can collect the image of the verification piece and send the collected verification piece to the host computer for image analysis. The host computer performs systematic verification based on the image of the verification piece to obtain a systematic inspection result, wherein the systematic inspection result includes the inspection result of whether the visual accuracy is abnormal, the inspection result of whether the imaging effect is abnormal, and the inspection result of whether the visual algorithm is abnormal.

[0267] The visual inspection system of this embodiment is adaptable to different application scenarios of visual inspection equipment. The device itself can have an automatic calibration function according to the characteristics of the measured object, and can automatically realize online calibration of the visual system light source, camera accuracy, and algorithm stability accuracy.

[0268] In the technical solution of the embodiment of the present application, the calibration parts are concentrated on the mounting plate, and the movement of the calibration parts is achieved by driving the mounting plate, thereby adopting an automatic method to achieve fully automatic calibration, thereby solving the problem that the calibration time is long and the production capacity loss is large due to the influence of the operator's proficiency.

[0269] In this embodiment, the calibration piece moving device is used to move the calibration block to the calibration position so that the light source of the visual inspection device illuminates the calibration piece. The calibration piece moving device is used to move the calibration piece under the light source for visual inspection when calibration is required, thereby avoiding affecting the normal operation of the production line. When calibration is no longer required, the calibration piece is automatically retracted. When calibration is required, the calibration piece on the calibration piece moving device is moved under the light source for visual inspection, thereby achieving automated visual inspection.

[0270] The check digit is the position of visual inspection. The visual inspection device includes a light source, a visual inspection component, a processing device and a control device. The visual inspection component faces the check digit and is used to collect the image of the product on the check digit. The light source provides light for the check digit, the processing device is used to obtain the visual image, and the control device is used to realize the workflow control of the visual inspection device on the production line. By arranging visual inspection on the check digit, effective management and control of the production line can be achieved.

[0271] In this embodiment, when a system calibration is required using a calibration piece, the calibration piece is automatically moved to the calibration position, thereby achieving fully automatic calibration of the visual inspection device.

[0272] This embodiment concentrates the calibration parts on the mounting plate, and moves the calibration parts by driving the mounting plate, thereby realizing fully automatic calibration in an automatic manner, solving the problem of long calibration time and large loss of production capacity due to the influence of operator proficiency.

[0273] According to some embodiments of the present application, optionally, as shown in the flow diagram of the second embodiment of the visual inspection method in FIG6 , the verification piece includes a plurality of size verification parts with gradient-changing sizes and a color card verification part with gradient-changing grayscale values;

[0274] Step S10': the visual inspection device is used to obtain a first image of the size verification part and a second image of the color card verification part, and send the first image and the second image to the host computer;

[0275] In step S20 ′, the host computer is configured to determine a systematic verification result of the visual inspection device according to the first image and / or the second image.

[0276] Generally speaking, a product-like calibration piece is designed and manually placed on the inspection position of the piece to be tested to match the actual production process detection posture. Then, a camera is used to obtain the dimensional information of the grooves, protrusions, etc. on the calibration piece to see if it is within a reasonable error range with the predetermined dimensional information, thereby achieving camera accuracy calibration. This calibration method is only for camera accuracy calibration, which limits the calibration method.

[0277] To address the technical limitations of existing calibration methods, the present invention provides a calibration component with multiple dimensional calibration sections with gradient variations in size. This effectively implements accuracy calibration within the measuring range of 2D and 3D camera measuring instruments, ensuring that the linear offset of the measuring instruments is verified. Linear offset is a linear error or endpoint linearity. It measures the maximum deviation of the endpoints of the entire measuring range, that is, the deviation between the measured curve and the ideal straight line. It can also verify the accuracy error caused by distortion of the 2D camera lens. Based on the accuracy verification, further algorithm verification is performed, thereby achieving systematic verification of the visual inspection device.

[0278] It should be noted that, in order to facilitate real-time monitoring of products on the production line, a visual inspection device captures images of the products on the production line and analyzes the images to obtain product specifications and related parameters, such as product size and thickness, thereby achieving effective product monitoring. The products may be battery cells, battery packs, or other types of products, and this embodiment is not limited to this. In this embodiment, a battery cell is used as an example for illustration. Multiple inspection stations are provided on the production line for the production and processing of battery cells. Each inspection station is equipped with a visual inspection device. The visual inspection device can be a 2D camera, a 3D camera, a structured camera, an area scan camera, or a line scan camera, as well as a light source. The visual inspection device captures images of the verification piece and is also equipped with a light source to illuminate the verification piece, facilitating the visual inspection device to capture images of the verification piece. The verification piece can be placed on a transport line. When the visual inspection device detects the verification piece, it captures an image of the verification piece. Alternatively, the visual inspection device can be located on the transport line to automatically place the verification piece on the conveyor line when verification is required. Alternatively, the verification piece can be moved to the inspection station by a drive device, and this embodiment is not limited to this.

[0279] The visual inspection device includes a verification piece, which is used as a sample to achieve systematic inspection of the visual inspection device by collecting images of the verification piece. As shown in Figure 2, the schematic diagram of the verification piece and the verification part in the verification piece, the verification part includes a plurality of size verification parts with gradient changes in size and a color card verification part with gradient changes in grayscale values. The size verification part can be a gradient verification groove or protrusion, or other forms of size verification parts. The verification piece requires that the size of the groove or protrusion is periodic and covers the measuring range of the measuring tool as much as possible. The verification piece requires that the groove or protrusion is not a mirror surface, and the surface is frosted and rough. <Ra3.2。

[0280] Under normal circumstances, the dimensional information of grooves, protrusions, etc. on the calibration piece only has a gradient change in the depth dimension, while the width or length has no gradient change. There is a certain degree of randomness in the calibration of 2D cameras, and it is impossible to effectively verify and confirm the linear offset of the measuring range, and it is impossible to detect the accuracy impact caused by lens distortion. Therefore, this embodiment adds gradient changes in two-dimensional and three-dimensional dimensions to effectively realize the accuracy calibration within the measuring range of 2D cameras and 3D cameras, ensure that the linear offset of the measuring tool is verified, and at the same time, it can verify the accuracy error caused by distortion of the 2D camera lens.

[0281] In this embodiment, the calibration piece requires no less than 5 grooves or protrusions with gradient changes, and at the same time meets the 2D and 3D size gradient changes. It is suitable for the accuracy calibration of the measuring range of 2D cameras and 3D camera measuring tools, and is suitable for the linear offset analysis and calibration of 2D cameras and 3D camera measuring tools.

[0282] In a specific implementation, the first image is a captured image of a plurality of size verification parts with gradient changes in size. The verification part can also be set on the product to obtain an image of the product, and the visual algorithm is verified based on the image of the product. A visual inspection device is provided opposite the verification part. When the verification part is supplemented with light by a light source, the visual inspection device obtains an image of the size verification part on the verification part. The image of the size verification part is systematically analyzed to realize the detection of the visual inspection device. The second image is a captured image of the color card verification part with gradient changes in grayscale value. The first image and the second image can be a single image or multiple separate images. When there are multiple images, the first image and the second image of the corresponding area are captured by visual inspection devices arranged in different areas. When there is a single image, the image can be divided into regions to obtain the first image and the second image corresponding to the plurality of size verification parts with gradient changes in size and the color card verification part with gradient changes in grayscale value, thereby realizing visual inspection.

[0283] The calibration piece can also be placed on the product to obtain an image of the product, and the visual algorithm can be verified based on the product image. A visual inspection device is placed directly opposite the calibration piece. When the calibration piece is supplemented with light from a light source, the visual inspection device obtains an image of the dimensional verification part on the calibration piece. This image of the dimensional verification part is systematically analyzed to achieve detection by the visual inspection device.

[0284] In this embodiment, the systematic verification results include visual accuracy verification results, imaging effect verification results, and visual algorithm verification results, thereby achieving systematic verification of the visual inspection device.

[0285] This embodiment features multiple dimensional calibration sections with gradient-varying dimensions on the calibration component, effectively enabling accuracy calibration within the measuring range of 2D and 3D camera measuring instruments, ensuring calibration of the measuring instrument's linear offset. Linear offset is a measure of linear error or endpoint linearity. It measures the maximum deviation of the line formed by the endpoints across the entire measuring range, i.e., the deviation between the measured curve and the ideal line. This also verifies accuracy errors caused by 2D camera lens distortion. Furthermore, based on this accuracy verification, algorithm verification can be further performed, thereby achieving systematic calibration of the visual inspection device.

[0286] It should be noted that if any of the visual accuracy verification results, imaging effect detection results, and visual algorithm verification results show an abnormal result, the device to be tested is notified to stop working.

[0287] In this embodiment, the obtained parameters and comparison results are uploaded to the Manufacturing Execution System (MES), which then compares the structural parameters again. If any result is NG, the MES locks the machine, preventing the inspection system from continuing production. Simultaneously, the equipment alarms, prompting staff to conduct system troubleshooting and adjustments based on the NG items to promptly correct visual inspection errors. This ensures that the visual inspection system maintains high accuracy over time or after adjustments. For example, if an anomaly in visual accuracy is detected, the MES locks the machine, or if an anomaly in imaging or visual algorithm is detected, shuts down the machine for rectification, thereby improving effective production monitoring.

[0288] In order to ensure the safety of production line production and prevent defective products from leaving the factory, this embodiment promptly stops production for inspection when problems with visual accuracy, visual imaging, and visual algorithms on the production line are detected, thereby achieving effective management and control of production line monitoring.

[0289] When the system verification result is normal, the verification mode is determined to be ended and the relevant production equipment is notified to enter the production mode.

[0290] In this embodiment, when the visual accuracy verification results, imaging effect detection results and visual algorithm verification results are all normal, the visual inspection device interacts with the upstream workstation to inform the upstream workstation to start unloading and enter the production mode, thereby realizing automated management of production.

[0291] In this embodiment, in order to avoid the presence of other materials in the verification and other influencing factors in the production process during the production verification process, it is necessary to suspend the material transportation of the production line and enter the verification mode. However, after the verification is completed, production needs to be resumed as soon as possible and automatically restored to the production mode, thereby improving the automated management of the production line.

[0292] According to some embodiments of the present application, optionally, as shown in FIG7 , the visual detection method further includes:

[0293] Step S401: upon receiving the verification instruction, the host computer sends the verification instruction to the visual inspection device;

[0294] Step S402: Upon receiving the verification instruction, the visual inspection device notifies the workstation controller to stop conveying materials.

[0295] As shown in Figure 8, the overall process diagram of visual inspection, the manufacturing execution system creates a verification task, or the equipment sets the automatic verification time, triggers the visual inspection device to enter the automatic verification mode, and the visual inspection device interacts with the upstream station to inform the upstream station to stop unloading and enter the verification mode. The specific process is S301: determine whether there is material at the verification position. If not, execute S302: move the verification block to the verification position, take a picture with the camera, and perform picture analysis. S303: after completing the picture acquisition, the cylinder controls the verification block to return to its original position; S304: camera accuracy judgment: |measurement value-standard value|<T / 10, imaging effect verification: |measurement Value - standard value | < 10: If all are OK, execute S305: Algorithm verification: |Measurement value - standard value| < T / 10. If all are OK, execute S306: Upload to the manufacturing execution system. The visual system notifies the upstream to start unloading and enter the production mode. If one of the camera accuracy or performance effect fails, or the algorithm verification fails, execute S308: Upload to the manufacturing execution system, the equipment stops, and the alarm is sounded. S309: The staff conducts system troubleshooting and system correction, and returns to execute S302. If there is material at the check position, execute S307: Perform product inspection and perform verification after completion.

[0296] In this embodiment, in order to avoid other materials being checked and other influencing factors in the production process during the production verification, the material delivery to the production line is suspended, thereby improving the accuracy of the verification.

[0297] According to some embodiments of the present application, optionally, as shown in FIG9 , the visual detection method further includes:

[0298] In step S501, when the visual inspection device detects that there is no material on the verification position, it notifies the visual inspection device to move the verification piece to the verification position. The verification position is within the detection range of the visual inspection device so that the light source in the visual inspection device illuminates the verification piece.

[0299] In this embodiment, the driving device can be a cylinder or other devices that can realize the driving function. This embodiment does not limit this. Taking the cylinder as an example, the visual inspection device confirms that there is no product at the verification position through the perception of the sensor, controls the cylinder to extend, and moves the verification piece to the waiting position.

[0300] This embodiment moves the verification piece to the verification position through a driving device without manual operation of the verification piece, thereby realizing fully automatic verification and solving the problem that the verification time is long and the production capacity is greatly lost due to the influence of the operator's proficiency.

[0301] In step S502, when the visual inspection device detects that there is material on the check position, it continues to inspect the current material on the check position until the target material leaves the check position, and then notifies the driving device in the visual inspection device to move the check piece to the check position so that the light source in the visual inspection device illuminates the check piece.

[0302] In this embodiment, the target material is the last material in the current transport process. When the automatic verification mode is in progress, if material is detected at the check position, verification will be restarted when the last product is detected in order to complete the current process. For example, the visual inspection device uses a sensor to determine whether there is material at the check position. If so, the last product is inspected. If not, verification will begin.

[0303] The visual inspection device of this embodiment uses a sensor to determine whether there is material at the check position. If there is material, the last product inspection is performed, thereby realizing fully automatic verification and improving the efficiency of verification.

[0304] According to some embodiments of the present application, optionally, as shown in FIG10 , the visual detection method further includes:

[0305] Step S601 : When the visual inspection device detects the first image and / or the second image, it notifies the driving device in the visual verification device to move the verification piece back to its original position.

[0306] In a specific implementation, after the verification piece image is acquired, the control cylinder is triggered to retract to its original position, and the verification piece is moved to its original position.

[0307] After obtaining the verification image, this embodiment automatically moves the verification piece back to its original position through the driving device without the need for manual operation of the verification piece, thereby realizing fully automatic verification and solving the problem of long verification time and large production capacity loss that is easily affected by the operator's proficiency.

[0308] According to some embodiments of the present application, optionally, as shown in FIG11 , the visual detection method further includes:

[0309] Step S701: If the system check result is normal, the host computer notifies the station controller to start conveying materials;

[0310] Step S702: When the system check result is abnormal, the host computer notifies each detection device to perform shutdown detection and issues an alarm.

[0311] In the technical solution of the embodiment of the present application, the upper computer controls the production line according to the system verification result. When the system verification result is normal, the workstation controller is notified to start conveying materials, thereby improving the processing efficiency of the production line. When the system verification result is abnormal, each detection equipment is notified to perform shutdown detection and issue an alarm, thereby improving the effective monitoring of the production line.

[0312] According to some embodiments of the present application, optionally, as shown in FIG12 , the visual detection method further includes:

[0313] Step S801: The host computer performs a visual accuracy check based on the first image to obtain a visual accuracy check result;

[0314] Step S802: The host computer performs an imaging effect verification based on the second image to obtain an imaging effect verification result;

[0315] Step S803: If the visual accuracy verification result and the imaging effect verification result are both normal verification results, the host computer performs visual algorithm verification based on the target image to obtain a visual algorithm verification result;

[0316] In step S804 , the host computer determines the system verification result of the visual inspection device according to the visual accuracy verification result, the imaging effect verification result, and the visual algorithm verification result.

[0317] In this embodiment, the visual accuracy verification can be used to verify the shooting accuracy of the visual inspection device. The size of the verification part captured by the visual inspection device can be compared with the size of the actual verification part to obtain the visual accuracy verification result. For example, after analyzing the first image, the size of the size verification part on the verification part is 10mm, but the size of the actual size verification part is 9mm. Therefore, the visual accuracy verification is achieved through the size verification part in the first image, effectively achieving accuracy verification within the measuring range of 2D cameras and 3D cameras, ensuring that the linear offset of the measuring tool is verified. At the same time, the accuracy error caused by the distortion of the 2D camera lens can be verified to achieve accuracy verification. The color card verification part can be a colorimetric card. The colorimetric card can be a light gray, gray, and dark gray gradient standard color card, or an RGB three-primary color card. Black and white cameras are required to be equipped with light gray, gray, and dark gray gradient standard color cards, and color cameras are required to be equipped with RGB three-primary color cards for imaging effect verification. For example, after analyzing the colorimetric card of the first image, the grayscale value of the colorimetric card on the verification piece is 50, but the actual grayscale value of the colorimetric card is 60, so that the imaging effect verification is achieved through the colorimetric card in the first image.

[0318] The target image is an image that has been calibrated by a normal visual algorithm and retrieved from the image verification library. It can also be an image that has been calibrated by a normal visual algorithm and directly input. The current visual algorithm is used to identify the target image to obtain a recognition result, and then the recognition result is compared with the result that has been calibrated by the normal visual algorithm to determine whether the current visual algorithm is normal.

[0319] When the visual accuracy verification results and the imaging effect verification results are both normal verification results, the visual algorithm verification is performed to achieve a systematic verification result of the visual inspection device.

[0320] In this embodiment, the systematic verification result of the visual inspection device can also be determined based on the detection value and the standard value corresponding to the first image. By comparing the difference between the detection value and the standard value, the systematic verification of the visual inspection device is achieved. For example, when performing visual accuracy detection, the detection value can be the measured value of the structural parameter, and the standard value can be the parameter value of the actual detection object, so as to compare and achieve visual accuracy verification. When performing imaging effect detection, the detection value can be the grayscale value of the colorimetric card, and the standard value can be the grayscale value of the actual detection object, so as to compare and achieve imaging effect verification. When performing a visual algorithm, the detection value can be the measured value obtained by analyzing the structural parameters of the current visual algorithm, and the standard value can be the parameter value obtained by analyzing a good visual algorithm, so as to compare and achieve visual algorithm verification.

[0321] This embodiment obtains the systematic verification result of the visual inspection device by comparing the detection value with the standard value, thereby realizing the systematic verification result of the visual inspection device through quantified data and improving the accuracy of the systematic verification of the visual inspection device.

[0322] In this embodiment, the systematic verification result of the visual inspection device can also be determined based on the second image.

[0323] In this embodiment, in addition to the size calibration part, a colorimetric card is also provided on the calibration piece. Since the size calibration part can calibrate the visual accuracy and the colorimetric card can calibrate the imaging effect, the visual accuracy, imaging effect and visual algorithm can be verified at the same time, thereby realizing systematic calibration of the visual inspection device.

[0324] This embodiment acquires images through the size verification part on the verification piece to obtain visual accuracy verification results, detects the imaging effect through the colorimetric card on the verification piece, and simultaneously detects the visual algorithm, thereby achieving systematic verification of the visual inspection device.

[0325] Systematic verification of the visual inspection device can be achieved by comparing the differences between the detection value and the standard value. For example, when performing visual accuracy detection, the detection value can be the measured value of the structural parameter, and the standard value can be the parameter value of the actual detection object, so as to make a comparison and realize the verification of visual accuracy. When performing imaging effect detection, the detection value can be the grayscale value of the colorimetric card, and the standard value can be the grayscale value of the actual detection object, so as to make a comparison and realize the verification of imaging effect. When performing visual algorithm, the detection value can be the measured value obtained by analyzing the structural parameters of the current visual algorithm, and the standard value can be the parameter value obtained by analyzing a good visual algorithm, so as to make a comparison and realize the verification of the visual algorithm.

[0326] This embodiment obtains visual accuracy verification results by comparing the measured values ​​corresponding to the structural parameters with the standard values, and obtains imaging effect verification results by comparing the grayscale values ​​of the colorimetric card with the standard values, thereby realizing systematic verification of the visual inspection device through quantified data and improving the accuracy of systematic verification.

[0327] In a specific implementation, a systematic verification result of the visual inspection device is determined based on the detection value and the standard value corresponding to the first image and the detection value and the standard value corresponding to the second image.

[0328] This embodiment compares the structural parameter detection values ​​of the size verification part image with multiple gradient-changing sizes in the first image with standard values ​​to realize visual accuracy verification while performing visual algorithm verification, and compares the detection values ​​of the color card verification part with gradient-changing grayscale values ​​in the second image with standard values ​​to realize effectiveness verification while performing visual algorithm verification, thereby achieving a systematic verification result of the visual detection device.

[0329] In this embodiment, the visual accuracy verification result of the visual inspection device can be determined based on the measurement values ​​corresponding to the structural parameters.

[0330] In this embodiment, the structural parameters may be the size, depth, and thickness of the size verification part, and may also include other parameters. This embodiment does not limit this. The comparison is made based on the structural parameters of the verification piece. Since the structural parameter value can be a specific numerical value, the difference between the measured value and the actual value can be obtained more accurately. For example, taking the size of the size verification part as an example, in order to achieve visual accuracy detection, the image of the two-dimensional size verification part on the verification piece is analyzed to obtain the size of the size verification part with a length of 2 mm and a width of 1 mm, so that the visual accuracy verification is directly performed based on the measurement value analyzed in the image.

[0331] This embodiment obtains a visual accuracy verification result by comparing the measurement values ​​corresponding to the structural parameters, thereby achieving visual accuracy verification through quantified data and improving the accuracy of visual detection.

[0332] In order to obtain the visual accuracy verification result, a first difference between the measured value corresponding to the structural parameter and the preset standard value can also be determined; based on the first difference, the visual accuracy verification result of the visual inspection device is determined.

[0333] In this embodiment, the preset standard value may be the numerical value of the structural parameter of the actual detection object. According to the difference between the measured value corresponding to the structural parameter and the preset standard value, the measured value of the structural parameter obtained by image analysis and the parameter value of the actual detection object are obtained. For example, by analyzing the image of the two-dimensional size calibration part on the calibration piece, it is obtained that the length of the size calibration part is 2 mm, and the length of the actual detection object is 2.5 mm. Then, the measured value of the structural parameter obtained by image analysis differs from the parameter value of the actual detection object by 0.5 mm, so that the measured value of the structural parameter obtained by image analysis is inconsistent with the parameter value of the actual detection object. It can be seen that the image collected by the current visual detection device is different from the actual detection object. Therefore, the visual accuracy of the visual detection device is abnormal.

[0334] Similarly, when the length of the size calibration part is 2mm and the actual length of the inspection object is 2mm, the measured value of the structural parameter obtained by image analysis is consistent with the parameter value of the actual inspection object. It can be seen that the image collected by the current visual inspection device is the same as the parameter value of the actual inspection object. Therefore, the visual accuracy of the visual inspection device is normal.

[0335] This embodiment obtains the visual accuracy verification result by the difference between the measured value corresponding to the structural parameter and the standard value corresponding to the actual detection object, thereby effectively obtaining the error between the image analysis result and the actual result, and more accurately obtaining the accuracy of visual detection.

[0336] After obtaining a first difference between the measurement value corresponding to the structural parameter and the preset standard value, a visual accuracy verification result of the visual inspection device is determined according to the first difference and the first parameter threshold range.

[0337] In this embodiment, the first parameter threshold range is defined to determine whether the difference is within the allowable range. If the difference is within the first parameter threshold range, it is determined to be within the allowable range. If the difference is outside the first parameter threshold range, it is determined to be within the disallowed range, thereby providing a reasonable interval of allowable error to avoid misjudgment. For example, taking the parameter threshold range of less than 0.1mm as an example, by analyzing the image of the two-dimensional size verification part on the verification piece, it is obtained that the length of the size verification part is 2mm, and the length of the actual detection object is 2.5mm. The measured value of the structural parameter obtained by image analysis differs from the parameter value of the actual detection object by 0.5mm, which exceeds the parameter threshold range. It can be seen that the visual accuracy of the visual detection device is abnormal.

[0338] Similarly, taking the parameter threshold range of less than 0.1mm as an example, by analyzing the image of the two-dimensional size verification part on the verification piece, it is found that the length of the size verification part is 2mm, and the length of the actual detection object is 2.03mm. The measured value of the structural parameter obtained by image analysis differs from the parameter value of the actual detection object by 0.03mm, which does not exceed the parameter threshold range. It can be seen that there is no abnormality in the visual accuracy of the visual inspection device.

[0339] In a specific implementation, as shown in FIG13 , step S801 includes:

[0340] Step S805 : The host computer determines a first difference between a measurement value corresponding to a structural parameter of the size verification part in the first image and a preset standard value.

[0341] In step S806, the host computer obtains a verification result indicating abnormal visual accuracy of the visual inspection device when the first difference is greater than or equal to the first parameter threshold range. In step S807, the host computer obtains a verification result indicating normal visual accuracy of the visual inspection device when the first difference is less than the first parameter threshold range.

[0342] In this embodiment, the parameter threshold range can be flexibly adjusted according to actual needs to improve the flexibility of precision detection.

[0343] This embodiment obtains the visual accuracy verification result by comparing the difference between the measured value corresponding to the structural parameter and the standard value corresponding to the actual detection object with the parameter threshold range. Compared with comparison only by difference, by giving a certain threshold range, the detection result is considered abnormal only when it exceeds this range, thereby improving the accuracy of visual accuracy detection.

[0344] According to some embodiments of the present application, optionally, the method further includes:

[0345] Step S808: The host computer determines a first parameter threshold range according to the tolerance corresponding to the structural parameter.

[0346] In this embodiment, the tolerance is the variation corresponding to each structural parameter, which is equal to the absolute value of the algebraic difference between the maximum limit and the minimum limit. For example, for dimensions, the tolerance corresponding to the dimension parameter is 1 mm, and for areas, the tolerance corresponding to the dimension parameter is 1 mm². When comparing the measured value with the parameter value, if a fixed tolerance is used for comparison of the parameter threshold, comparison errors may occur. The first parameter threshold range can be less than T / 10, where T represents the tolerance corresponding to the structural parameter, that is, camera accuracy verification: |measured value - standard value| < T / 10. For example, when performing the verification on the dimension in the structural parameters, if the measured value of the structural parameter obtained by image analysis differs from the parameter value of the actual detection object by 0.03 mm, then the parameter threshold range corresponding to the dimension parameter is compared with less than 0.1 mm, avoiding comparison with the parameter threshold range corresponding to the area parameter of less than 1 mm², thus preventing misjudgment situations.

[0347] Before obtaining the parameter threshold range in this embodiment, the parameter threshold range is determined according to the corresponding structural parameter. Since the standards corresponding to different structural parameters are different, the parameter threshold range is determined by the tolerance corresponding to the parameter, making the parameter threshold range adaptable to the structural parameter, thereby improving the rationality of visual accuracy detection.

[0348] According to some embodiments of the present application, as shown in FIG. 14, optionally, the visual detection method further includes:

[0349] Step S809, the host computer detects the first image through the target vision algorithm to obtain the target standard value corresponding to the structural parameter.

[0350] Step S810, the host computer determines the second difference between the measured value of the structural parameter of the dimension verification part in the first image and the target standard value.

[0351] Step S811, when the second difference is greater than or equal to the second parameter threshold range, the host computer obtains the verification result that the vision algorithm of the visual detection device is abnormal.

[0352] Step S812, when the second difference is less than the second parameter threshold range, the host computer obtains the verification result that the vision algorithm of the visual detection device is normal.

[0353] It should be noted that the host computer can determine the first vision algorithm verification result of the visual detection device according to the measured value corresponding to the structural parameter.

[0354] In this embodiment, the measurement value corresponding to the structural parameter is compared with the measurement value obtained by the good visual analysis algorithm to realize the verification of the visual algorithm. For example, by analyzing the image of the two-dimensional size verification part on the verification piece, the length of the size verification part is obtained to be 2 mm. The measurement value 1.8 mm is obtained by analyzing the image of the two-dimensional size verification part on the verification piece using the good visual analysis algorithm, thereby realizing the verification of the visual algorithm.

[0355] When the measurement value of the size calibration part obtained by analyzing the image of the two-dimensional size calibration part on the calibration piece is consistent with the measurement value obtained by analyzing the image of the two-dimensional size calibration part on the calibration piece through a good visual analysis algorithm, it is determined that the current visual algorithm is normal. When the measurement value of the size calibration part obtained is inconsistent with the measurement value obtained by analyzing the image of the two-dimensional size calibration part on the calibration piece through a good visual analysis algorithm, it is determined that the current visual algorithm is abnormal, thereby realizing the detection of the visual algorithm.

[0356] In addition to performing visual accuracy testing through the measurement values ​​corresponding to the structural parameters, this embodiment can also perform visual algorithm verification, thereby achieving systematic visual verification and improving the comprehensiveness and effectiveness of production line monitoring.

[0357] The host computer can also determine a second difference between the measured value corresponding to the structural parameter and the target standard value; and determine the first visual algorithm verification result of the visual detection device based on the second difference.

[0358] In this embodiment, the target standard value may be a measurement value obtained by detection through a good visual algorithm. According to the difference between the measurement value corresponding to the structural parameter and the target standard value, the difference between the measurement value of the structural parameter obtained by image analysis and the measurement value obtained by detection through the good visual algorithm is obtained. For example, by analyzing the image of the two-dimensional size calibration part on the calibration piece, the length of the size calibration part is 2 mm, and the measurement value obtained by analyzing the image of the two-dimensional size calibration part on the calibration piece through the good visual analysis algorithm is 2.5 mm. The measurement value of the structural parameter obtained by image analysis and the measurement value obtained by analyzing the image of the two-dimensional size calibration part on the calibration piece through the good visual analysis algorithm differ by 0.5 mm. Therefore, the measurement value of the structural parameter obtained by image analysis is inconsistent with the measurement value obtained by analyzing the image of the two-dimensional size calibration part on the calibration piece through the good visual analysis algorithm. It can be seen that there is a difference in the detection results between the current visual algorithm and the good visual algorithm. Therefore, there is an abnormality in the current visual algorithm.

[0359] Similarly, when the length of the dimensional verification part is 2 mm, and the measured value obtained by analyzing the image of the two-dimensional dimensional verification part on the verification piece through a good vision analysis algorithm is 2 mm, the measured value of the structural parameters obtained by image analysis is consistent with the measured value obtained by analyzing the image of the two-dimensional dimensional verification part on the verification piece through a good vision analysis algorithm. It can be seen that the detection results of the current vision algorithm and the good vision algorithm are the same. Therefore, the current vision algorithm is normal.

[0360] In this embodiment, the measured value corresponding to the structural parameter is compared with the measured value of the structural parameter analyzed by the normal vision detection algorithm, so as to obtain the difference between the analysis result of the current vision algorithm and the analysis result detected by the normal vision detection algorithm. Therefore, the difference between the current vision algorithm and the normal vision detection algorithm is obtained, thus realizing the verification of the vision algorithm.

[0361] In this embodiment, the target vision algorithm is a good vision algorithm that has been verified. To facilitate comparison with the measured value of the current vision algorithm, the dimensional verification part is analyzed by the target vision algorithm at the same time to obtain the target standard value corresponding to the structural parameter, so as to realize the comparison with the current measured value. For example, the measured value obtained by analyzing the image of the two-dimensional dimensional verification part on the verification piece through a good vision analysis algorithm is 2 mm, and this measured value of 2 mm is used as the target standard value. Thus, different vision algorithms are used to analyze the same analysis object, avoiding the influence of using different analysis objects for comparison on the accuracy of the comparison result. In this embodiment, to more effectively obtain the difference between the current vision algorithm and the normal vision detection algorithm, the dimensional verification part is analyzed by the target vision algorithm to obtain the target standard value corresponding to the structural parameter, so as to be comparable with the measured value corresponding to the structural parameter, realizing the standard unity of the comparison between the current vision algorithm and the normal vision detection algorithm, and improving the accuracy of the vision algorithm verification.

[0362] In this embodiment, the second parameter threshold range is defined to determine whether the difference is within the allowable range. If the difference is within the second parameter threshold range, it is determined that it is within the allowable range; if the difference is outside the second parameter threshold range, it is determined that the difference is outside the allowable range, thus providing a reasonable interval of allowable error to avoid misjudgment. The second parameter threshold range can be T / 10, where T represents the tolerance corresponding to the detection object, that is, for vision algorithm verification: |measured value - standard value| < T / 10. For example, taking the parameter threshold range being less than 0.1 mm as an example, by analyzing the image of the two-dimensional dimensional verification part on the verification piece, the length of the dimensional verification part is obtained as 2 mm, and the length detected by the good vision algorithm is 2.5 mm. Then the difference between the measured value of the structural parameter obtained by image analysis and the parameter value detected by the good vision algorithm is 0.5 mm, which exceeds the parameter threshold range. It can be seen that the vision algorithm of the vision detection device is abnormal.

[0363] Similarly, taking the parameter threshold range of less than 0.1mm as an example, by analyzing the image of the two-dimensional size verification part on the verification piece, it is found that the length of the size verification part is 2mm, and the length obtained by detecting with a good visual algorithm is 2.03mm. The measured value of the structural parameter obtained by image analysis differs from the parameter value obtained by detecting with a good visual algorithm by 0.03mm, which does not exceed the parameter threshold range. It can be seen that there is no abnormality in the visual algorithm of the visual detection device.

[0364] In specific implementations, the parameter threshold range can also be flexibly adjusted according to actual needs to improve the flexibility of precision detection.

[0365] As shown in Figure 15, the visual system verification process flow diagram is as follows: S801': Move the product to the verification position, triggering the camera to capture the product image; S802': Pass the captured image into the image processing and analysis module, namely the visual inspection algorithm; S803': Based on the output of the visual inspection algorithm, the equipment compares it with the set rules to determine whether the product is defective; S804': The result is uploaded to the Manufacturing Execution System (MES), and the Manufacturing Execution System automatically records the product status information based on the result, and the equipment performs the preset action based on the result.

[0366] A manufacturing execution system (MES) is an industrial production management system used to optimize and monitor the entire manufacturing process, from raw materials to final product. MES coordinates the planning, execution, control, and monitoring of all aspects of production activities by interacting with various devices, systems, and personnel. It is often integrated with other enterprise information systems, such as enterprise resource planning (ERP), product lifecycle management (PLM), and supply chain management (SCM), to achieve more efficient, flexible, and visualized production management.

[0367] As can be seen from the process, if the first step of image acquisition is normal during production, that is, the hardware such as the camera and light source is functioning properly and the captured images are stable, this part of the design includes camera accuracy verification and imaging verification. The normal operation of the visual algorithm will affect the determination of whether the product is defective. The algorithm can be designed to determine whether the specifications have changed and require verification.

[0368] Schematic diagram of the process for visual algorithm verification as shown in Figure 16. Algorithm verification: S801": Replace the product pictures collected by the current camera with pictures in the algorithm verification library; S802": Input the collected pictures into the image processing and analysis module, that is, the visual detection algorithm; S803": According to the output of the visual detection algorithm result, compare the device with the set specifications to determine whether the product is defective; S804": Compare and verify the current detection result with the standard result, and determine whether the algorithm and specifications are normal according to whether it is within the standard range; S805": Upload the verification result to the manufacturing execution system, and the device executes preset actions according to the verification result; S806": According to the number N of pictures in the picture library, loop through and execute N times.

[0369] It can be seen from the process that the algorithm verification picture library is determined manually, and its result is known. When detecting with the already confirmed and improved algorithm and standard specifications, its result is the standard. Then, when performing algorithm verification, when detecting the product pictures with known results, after being processed by the algorithm and specifications, the results should be the same. If they are not the same, there is a problem with the algorithm, as shown in the detection data in Table 1:

[0370] Table 1

[0371] The verification process is as follows: According to the detection result, |measured value - standard value| < T / 10. According to the specifications, T = 11 - 10 = 1. At this time, the detection results of pictures numbered 1, 2, and 3 are within the reasonable range, while the detection result of picture numbered 4 has a deviation of 0.2 and is not within the reasonable range. Therefore, the algorithm stability is NG, and there is a problem with the measurement system.

[0372] According to the specification verification, the original specification is 10 - 11, and the current specification is 10.5 - 11. Among them, the original rule was determined when establishing the standard, and the current rule is the rule read from the current device settings. The two are inconsistent, indicating that the device specifications have been maliciously modified and there is a problem with the measurement system, thereby realizing the verification of the visual algorithm.

[0373] In this embodiment, the difference between the measured value corresponding to the structural parameter and the standard value corresponding to the normal visual detection algorithm is compared with the parameter threshold range to obtain the visual algorithm verification result, and by comparing with the specifications, the systematic verification of the visual detection device is further realized. Compared with only comparing through the difference, by giving a certain threshold range, it is considered that the visual algorithm is abnormal only when exceeding this range, thereby improving the accuracy of the visual algorithm verification.

[0374] According to some embodiments of the present application, optionally, as shown in Figure 17, step S802 includes:

[0375] Step S813, the host computer determines the third difference between the gray value of the color card verification part in the second image and the preset gray value.

[0376] Step S814: When the third difference is greater than or equal to the third parameter threshold range, the host computer determines that the imaging effect of the visual inspection device is abnormal.

[0377] In step S815 , when the third difference is less than the third parameter threshold range, the host computer determines that the imaging effect of the visual inspection device is normal.

[0378] In this embodiment, grayscale value is an image representation concept used in digital image processing technology. Grayscale value refers to the brightness value of a pixel in a grayscale digital image, representing the depth of the color in the digital image. Grayscale levels range from 0 to 255, with white at 255 and black at 0. Grayscale images can display different shades of any color, even different colors at different brightnesses. Comparisons based on the grayscale values ​​of a colorimetric chart can be more accurate because the grayscale values ​​of the colorimetric chart can be specific numerical values. For example, to verify imaging effects, if, after analyzing the colorimetric chart in the second image, the grayscale value of the colorimetric chart on the verification piece is determined to be 50, the imaging effect can be verified directly based on the grayscale values ​​analyzed in the image.

[0379] This embodiment obtains the imaging effect verification result by comparing the grayscale value of the color card, thereby realizing the verification of the imaging effect through quantitative data and improving the accuracy of the imaging effect detection.

[0380] In this embodiment, the preset grayscale value may be the grayscale value of the actual detection object, and image analysis is performed on the colorimetric card to obtain the difference between the corresponding measurement value and the preset standard value, thereby obtaining the measurement value of the colorimetric card obtained by image analysis and the grayscale value of the actual detection object. For example, after analyzing the colorimetric card of the second image, the grayscale value of the colorimetric card on the calibration piece is 50, and the grayscale value of the same position of the actual colorimetric card is 60. The measurement value obtained by the image analysis differs from the parameter value of the actual detection object by 10, thereby obtaining that the current measurement value obtained by the image analysis is inconsistent with the parameter value of the actual detection object. It can be seen that the image captured by the current visual detection device is different from the actual detection object. Therefore, the imaging effect of the visual detection device is abnormal.

[0381] Similarly, after analyzing the colorimetric card of the second image, the grayscale value of the colorimetric card on the calibration piece is 50, and the grayscale value of the same position on the actual colorimetric card is 50. The measurement value obtained by the image analysis is consistent with the parameter value of the actual detection object. It can be seen that the image collected by the current visual detection device is the same as the parameter value of the actual detection object. Therefore, the imaging effect of the visual detection device is normal.

[0382] This embodiment obtains the imaging effect verification result by the difference between the grayscale value and the grayscale value corresponding to the actual detection object, thereby effectively obtaining the error between the image analysis result and the actual result, and more accurately realizing the detection of the imaging effect.

[0383] In this embodiment, the third parameter threshold range is defined to determine whether the grayscale value difference is within the allowable range. If the difference is within the third parameter threshold range, it is determined to be within the allowable range. If the difference is outside the third parameter threshold range, it is determined to be within the disallowed range, thereby providing a reasonable interval of a certain allowable error to avoid misjudgment. The third parameter threshold range can be 10, that is, imaging effect verification: |measurement value-standard value|<10. For example, taking the parameter threshold range of less than 6 as an example, after analyzing the colorimetric card of the second image, the grayscale value of the colorimetric card on the verification piece is 50, and the grayscale value of the same position on the actual colorimetric card is 60. The measurement value obtained by the image analysis differs from the parameter value of the actual detection object by 10, which exceeds the parameter threshold range. It can be seen that the imaging effect of the visual detection device is abnormal.

[0384] Similarly, taking the parameter threshold range of less than 6 as an example, after analyzing the colorimetric card of the second image, the grayscale value of the colorimetric card on the calibration piece is 50, and the grayscale value of the same position on the actual colorimetric card is 55. The measurement value obtained by image analysis differs from the parameter value of the actual detection object by 5, which does not exceed the parameter threshold range. It can be seen that there is no abnormality in the imaging effect of the visual detection device.

[0385] In specific implementations, the parameter threshold range can also be flexibly adjusted according to actual needs to improve the flexibility of precision detection.

[0386] This embodiment obtains the imaging effect verification result by comparing the difference between the current grayscale value and the standard value corresponding to the actual detection object with the parameter threshold range. Compared with comparison only by difference, by giving a certain threshold range, the detection result is considered abnormal only when it exceeds the range, thereby improving the accuracy of imaging effect detection.

[0387] According to some embodiments of the present application, optionally, as shown in FIG18 , the visual detection method further includes:

[0388] In step S816, the host computer detects the second image using a target vision algorithm to obtain a target grayscale value.

[0389] In step S817, the host computer determines a fourth difference between the grayscale value of the color card verification portion in the second image and the target grayscale value.

[0390] Step S818: When the fourth difference is greater than or equal to the fourth parameter threshold range, the host computer obtains a verification result indicating that the visual algorithm of the visual detection device is abnormal.

[0391] In step S819, when the fourth difference is less than the fourth parameter threshold range, the host computer obtains a verification result indicating that the visual algorithm of the visual inspection device is normal.

[0392] In this embodiment, the grayscale value obtained by the current visual algorithm analysis is compared with the grayscale value obtained by the good visual analysis algorithm to achieve verification of the visual algorithm. For example, after analyzing the colorimetric card of the second image, the grayscale value of the colorimetric card on the verification piece is 50. The grayscale value of the same position of the colorimetric card on the verification piece is calibrated to 60 by the good visual analysis algorithm, thereby achieving verification of the visual algorithm.

[0393] When the measurement value of the size verification part is obtained by analyzing the image of the two-dimensional size verification part on the verification part, and the measurement value is consistent with the measurement value obtained by analyzing the image of the two-dimensional size verification part on the verification part through a good visual analysis algorithm, it is determined that the current visual algorithm is normal. When the measurement value of the size verification part is inconsistent with the measurement value obtained by analyzing the image of the two-dimensional size verification part on the verification part through a good visual analysis algorithm, it is determined that the current visual algorithm is abnormal, thereby realizing the detection of the visual algorithm.

[0394] In addition to detecting the imaging effect through grayscale values, this embodiment can also verify the visual algorithm, thereby achieving systematic visual verification and improving the comprehensiveness and effectiveness of production line monitoring.

[0395] In this embodiment, the target grayscale value may be a measurement value detected by a good vision algorithm. According to the difference between the grayscale value of the colorimetric card and the target standard value, the difference between the measurement value obtained by image analysis and the measurement value obtained by detection by the good vision algorithm is obtained. For example, after analyzing the colorimetric card of the second image, the grayscale value of the colorimetric card on the calibration piece is 50, and the grayscale value of the same position of the colorimetric card on the calibration piece after calibration by the good vision analysis algorithm is 60. The parameter value of the measurement value obtained by image analysis and the measurement value obtained by analyzing the image of the colorimetric card on the calibration piece by the good vision analysis algorithm differs by 10, so the measurement value obtained by image analysis is inconsistent with the measurement value obtained by analyzing the image of the two-dimensional size calibration part on the calibration piece by the good vision analysis algorithm. It can be seen that there is a difference in the detection results between the current vision algorithm and the good vision algorithm. Therefore, there is an abnormality in the current vision algorithm.

[0396] Similarly, after analyzing the color comparison card, the grayscale value of the color comparison card on the calibration piece is 50. The grayscale value of the same position of the color comparison card on the calibration piece is calibrated with a good visual analysis algorithm and is 50. The measurement value obtained by image analysis is consistent with the measurement value obtained by analyzing the color comparison card on the calibration piece with the good visual analysis algorithm. It can be seen that the detection results of the current visual algorithm are the same as those of the good visual algorithm. Therefore, the current visual algorithm is normal.

[0397] This embodiment compares the grayscale value with the grayscale value analyzed by the normal visual detection algorithm to obtain the difference between the analysis result of the current visual algorithm and the analysis result detected by the normal visual detection algorithm, thereby obtaining the difference between the current visual algorithm and the normal visual detection algorithm, thereby realizing the verification of the visual algorithm.

[0398] In this embodiment, the target vision algorithm is a good vision algorithm that has been calibrated. In order to facilitate comparison with the measurement value of the current vision algorithm, the colorimetric card is analyzed by the target vision algorithm to obtain the target standard value corresponding to the colorimetric card, thereby realizing comparison with the current measurement value. For example, the grayscale value of the same position of the colorimetric card on the calibration piece is calibrated to 50 by a good vision analysis algorithm, and the grayscale value of 50 is used as the target standard value, so that different vision algorithms are used for analysis based on the same analysis object, avoiding the use of different analysis objects for comparison, which affects the accuracy of the comparison results.

[0399] In order to more effectively obtain the difference between the current visual algorithm and the normal visual detection algorithm, this embodiment uses the target visual algorithm to analyze the color comparison card to obtain the target standard value corresponding to the color comparison card, which can be compared with the grayscale value obtained by the current visual algorithm analysis, thereby achieving the standard unification of the comparison between the current visual algorithm and the normal visual detection algorithm, and improving the accuracy of the visual algorithm verification.

[0400] In this embodiment, the fourth parameter threshold range is defined to determine whether the grayscale value difference is within the allowable range. If the difference is within the fourth parameter threshold range, it is determined to be within the allowable range. If the difference is outside the fourth parameter threshold range, it is determined to be within the unallowable range, thereby providing a reasonable interval of a certain allowable error to avoid misjudgment. The fourth parameter threshold range can be 10, that is, the visual algorithm verification: |measurement value-standard value|<10. For example, taking the parameter threshold range as less than 10 as an example, after analyzing the colorimetric card, the grayscale value of the colorimetric card on the calibration piece is 50. The grayscale value of the same position of the colorimetric card on the calibration piece is 60 after calibration by a good visual analysis algorithm. The measurement value obtained by image analysis differs from the parameter value obtained by good visual algorithm detection by 10, which exceeds the parameter threshold range. It can be seen that the visual algorithm of the visual detection device is abnormal.

[0401] Similarly, taking the parameter threshold range of less than 10 as an example, after analyzing the color comparison card, the grayscale value of the color comparison card on the calibration piece is 50. The grayscale value of the same position of the color comparison card on the calibration piece is 55 when calibrated by a good visual analysis algorithm. The measurement value obtained by image analysis differs from the parameter value obtained by detection by a good visual algorithm by 5, which does not exceed the parameter threshold range. It can be seen that there is no abnormality in the visual algorithm of the visual inspection device.

[0402] In specific implementations, the parameter threshold range can also be flexibly adjusted according to actual needs to improve the flexibility of precision detection.

[0403] This embodiment obtains the visual algorithm verification result by comparing the difference between the grayscale value corresponding to the current visual algorithm and the standard value corresponding to the normal visual detection algorithm with the parameter threshold range. Compared with comparison only by difference, by giving a certain threshold range, the visual algorithm is considered abnormal only when it exceeds this range, thereby improving the accuracy of visual algorithm verification.

[0404] According to some embodiments of the present application, optionally, as shown in FIG19 , before step S803, the following steps are further included:

[0405] In step S820, the host computer obtains the first image, the second image, the third image, and the fourth image with calibrated parameter values, wherein the parameter value of the first image is within the first range, the parameter value of the second image is within the second range, the parameter value of the third image is within the third range, and the parameter value of the third image is within the fourth range, the first range and the second range are different, and the third range and the fourth range are different; an image verification library is established based on the first image, the second image, the third image, and the fourth image. In this embodiment, the first image can be a normal image, the second image can be a normal limit image, the third image can be an abnormal image, and the fourth image can be an abnormal limit image. The product image verification library must have NG products, NG limit samples, OK products, and OK limit samples. Each defect of NG products and NG limit samples must be no less than 3EA, and each defect of OK products and OK limit samples must be no less than 3EA. The verification library is established by collecting images of typical defective products produced during production, or artificially creating typical defects, and then taking pictures and storing the images with a camera.

[0406] This embodiment uses a pre-established image verification library to detect the visual algorithm, which is more efficient than directly comparing the verification results.

[0407] In a specific implementation, the host computer numbers the first picture, the second picture, the third picture, and the fourth picture; and establishes a picture verification library according to the numbered first picture, the second picture, the third picture, and the fourth picture.

[0408] In this embodiment, the pictures are numbered and analyzed using a complete algorithm to obtain the results, which are set as standard values. The numbering method can be through coding identification or other methods. This embodiment does not limit this. By numbering, the positioning and tracking of pictures can be achieved, which facilitates gallery management.

[0409] In this embodiment, when establishing the picture verification library, the picture samples are managed by numbering, so as to achieve the effective management of the picture verification library, but it is not convenient for the subsequent adjustment and update of the picture verification library.

[0410] According to some embodiments of the present application, optionally, it further includes:

[0411] Step S821, the host computer uses the target vision detection algorithm to evaluate the sample image, and obtains the first picture, the second picture, the third picture and the fourth picture with calibrated parameter values.

[0412] In this embodiment, the pictures in the verification library are analyzed by a complete algorithm to obtain the results, which are set as the standard values. For example, for picture A, the length in the structural parameters obtained by good vision algorithm analysis is 5 mm, so as to facilitate the verification of the vision algorithm.

[0413] In this embodiment, the sampled pictures are pre-evaluated by a complete vision detection algorithm, and the evaluated parameter values are used as the standard values for subsequent comparison, so as to achieve the verification of the vision algorithm.

[0414] According to some embodiments of the present application, optionally, as shown in FIG. 20, step S803 includes:

[0415] Step S818', select the target image from the picture verification library.

[0416] Step S819', perform image recognition on the target image through the vision detection device to obtain the parameter values of the target image.

[0417] Step S820', compare the parameter values with the calibrated parameter values corresponding to the target image.

[0418] Step S821', when the difference between the parameter values and the calibrated parameter values corresponding to the target image exceeds the fifth parameter threshold range, determine that the vision algorithm is abnormal.

[0419] In this embodiment, after the vision system completes the accuracy and imaging verification results, the vision system performs algorithm verification. The vision algorithm traverses the algorithm verification library, and obtains the result parameters of the verification pictures after algorithm analysis. If the obtained result parameters are the same as the predetermined standard parameters or within the reasonable error range of the predetermined standard parameters, it means that the algorithm accuracy is good, and the algorithm verification: |measured value - standard value| < T / 10, where T is the tolerance of the detection object. If the obtained result parameters deviate greatly from the predetermined standard parameters, it means that there are problems such as algorithm parameter, detection specification optimization iteration or human illegal change in the vision algorithm, resulting in algorithm abnormality, and the algorithm is in a failure working state, and the algorithm accuracy is abnormal.

[0420] According to some embodiments of the present application, as shown in FIG. 21, a specific implementation manner is provided:

[0421] Step S900, the manufacturing execution system creates a verification task, or the device sets an automatic verification time to trigger the vision detection device to enter the automatic verification mode.

[0422] Step S901, the vision detection device interacts with the upstream station to inform the upstream station to stop blanking and enter the verification mode.

[0423] Step S902, the vision detection device determines whether there is material at the verification position through a sensor;

[0424] Step S903, if there is material, the last product is detected. If there is no material, verification starts.

[0425] Step S904, the vision detection device confirms that there is no product at the verification position through the perception of the sensor, controls the cylinder to extend, and moves the verification piece to the material waiting position.

[0426] Step S905, after the camera performs imaging analysis on the dimension verification part on the verification piece to obtain the parameters of the dimension verification part, the camera accuracy verification obtains the dimension parameters of the groove or protrusion, and the imaging effect verification obtains the gray value of the color card.

[0427] Step S906, after obtaining the verification piece picture, trigger the control cylinder to retract to its original position and move the verification piece back to its original position.

[0428] Step S907, compare the obtained structural parameters with the predetermined structural parameters. If the obtained structural parameters are the same as the predetermined structural parameters or within the reasonable error range of the predetermined structural parameters, it means that the obtained groove or protrusion structural parameters have not distorted, the lens has not loosened or defocused, etc., the structural parameters of the color card have not mutated, the light source brightness, camera aperture, camera exposure, gain, etc. have not mutated, the accuracy and imaging quality of the camera are good, and the camera and the light source are in normal working states (camera accuracy verification: |measured value - standard value| < T / 10, imaging effect verification: |measured value - standard value| < 10, where T is the tolerance of the detection object); if the deviation between the obtained groove or protrusion structural parameters and the predetermined structural parameters is large, it means that the obtained structural parameters have distorted, or the lens has loosened or defocused, etc., the accuracy of the camera has deteriorated, and the camera is in an abnormal working state; if the deviation between the obtained gray value of the color card and the predetermined gray value is large, it means that the light source brightness, camera aperture, camera exposure, gain, etc. have mutated, the imaging effect is poor, and the imaging system (camera, light source) is in an abnormal working state. After the vision detection device completes the accuracy and imaging verification results, the vision detection device performs algorithm verification.

[0429] Step S908: The vision algorithm traverses the algorithm verification library, and after algorithm analysis, obtains the result parameters of the verification picture. If the obtained result parameters are the same as the predetermined standard parameters or within the reasonable error range of the predetermined standard parameters, it indicates that the algorithm accuracy is good, and the algorithm verification is: |measured value - standard value| < T / 10, where T is the tolerance of the detection object. If the obtained result parameters deviate greatly from the predetermined standard parameters, it indicates that there are problems such as algorithm parameter, detection specification optimization iteration, or human violation change in the vision algorithm, resulting in algorithm abnormality, and the algorithm is in a failure working state with abnormal algorithm accuracy.

[0430] Step S909: Upload the obtained various parameters and comparison results to the manufacturing execution system. The manufacturing execution system compares the structural parameters again. If any result is NG, the manufacturing execution system locks the machine to prevent the detection system from producing. At the same time, the device alarms.

[0431] Step S910: The staff needs to conduct system troubleshooting and adjustment based on the NG items to promptly correct the vision detection error. Ensure that the vision detection system still has good accuracy after a long time or adjustment.

[0432] Step S911: The vision detection device interacts with the upstream station to inform the upstream station to start blanking and enter the production mode.

Claims

1. A visual inspection system, wherein: The visual inspection system comprises: A visual inspection device, comprising a verification piece and a verification piece moving device, wherein the verification piece moving device comprises a movably mounted mounting plate, wherein the mounting plate can be located at a verification position during the movable travel of the mounting plate, wherein the verification position is within the detection range of the visual inspection device, and wherein the verification piece is arranged on the mounting plate; A visual inspection device, used to obtain an image of the verification piece and send the image of the verification piece to a host computer; The host computer is used to determine the systematic verification result of the visual inspection device according to the image of the verification part.

2. The visual inspection system according to claim 1, wherein: The verification part includes a plurality of size verification parts with gradient-changing sizes and a color card verification part with gradient-changing grayscale values; The visual inspection device is further used to obtain a first image of the size verification unit and a second image of the color card verification unit, and send the first image and the second image to a host computer; The host computer is further used to determine a systematic verification result of the visual inspection device according to the first image and / or the second image.

3. The visual inspection system according to claim 2, wherein: The visual verification device also includes: a support base; and, The driving device comprises a fixed part and a movable part which can move linearly relative to each other, the fixed part is arranged on the supporting base, and the movable part is connected to the mounting plate.

4. The visual inspection system according to claim 3, wherein: The visual inspection device also includes a sliding guide structure, which includes a slide rail and a slider that slide with each other, and the slide rail and the slider are arranged between the support base and the mounting plate.

5. The visual inspection system according to claim 4, wherein: The visual inspection device also includes an adapter, two of the slide rails are provided, and two corresponding slide blocks are provided; One group of the slide rails and the sliding blocks that cooperate with each other is arranged between the support base and the adapter seat, and the other group of the slide rails and the sliding blocks is arranged between the mounting plate and the adapter seat.

6. The visual inspection system according to claim 1, wherein: The verification piece includes: a plate body having a calibration surface; A plurality of dimension verification parts are arranged on the verification surface at intervals along a straight line direction, and the plurality of dimension verification parts have a length dimension in a first direction and a width dimension in a second direction, and the length dimensions of the plurality of dimension verification parts vary in a gradient, and / or the width dimensions of the plurality of dimension verification parts vary in a gradient, wherein the first direction and the second direction are directions perpendicular to each other in a horizontal plane.

7. The visual inspection system of claim 6, wherein: The size checking portion includes a protrusion or a groove formed on the plate body.

8. The visual inspection system according to claim 6, wherein: The sizes of the plurality of size verification parts along a third direction vary in a gradient manner, and the third direction is arranged perpendicular to both the first direction and the second direction.

9. The visual inspection system according to claim 6, wherein: It also includes a color card calibration part which is arranged on the calibration surface of the board body and has a grayscale value that changes in a gradient.

10. The visual inspection system according to claim 6, wherein: The material of the plate body includes aluminum alloy; and / or, The roughness of the verification surface of the plate body is less than a preset roughness threshold.

11. A visual inspection method, wherein: The visual inspection system includes a visual inspection device, a visual verification device and a host computer, wherein the visual verification device includes a verification piece and a verification piece moving device, wherein the verification piece moving device includes a movably mounted mounting plate, wherein the mounting plate can be in a verification position during the movable travel of the mounting plate, wherein the verification position is within the detection range of the visual inspection device, and wherein the verification piece is arranged on the mounting plate; The visual detection method comprises: The visual inspection device obtains the image of the verification piece and sends the image of the verification piece to the host computer; The host computer determines the systematic verification result of the visual inspection device according to the image of the verification part.

12. The visual inspection method according to claim 11, wherein: The verification piece includes a plurality of size verification parts with gradient changes in size and a color card verification part with gradient changes in grayscale value, and the visual inspection method further includes: The visual inspection device obtains a first image of the size verification unit and a second image of the color card verification unit, and sends the first image and the second image to a host computer; The host computer determines a systematic verification result of the visual inspection device according to the first image and / or the second image.

13. The visual inspection method according to claim 11, wherein: The visual detection method also includes: When receiving the verification instruction, the host computer sends the verification instruction to the visual inspection device; When receiving the verification instruction, the visual inspection device notifies the station controller to stop conveying materials.

14. The visual inspection method according to claim 13, wherein: The visual detection method also includes: When the visual inspection device detects that there is no material on the verification position, the driving device in the visual inspection device is notified to move the verification piece to the verification position, and the verification position is within the detection range of the visual inspection device; When the visual inspection device detects that there is material on the check position, it continues to inspect the current material on the check position until the target material leaves the check position, and then notifies the driving device in the visual inspection device to move the check piece to the check position.

15. The visual inspection method according to claim 12, wherein: The visual detection method also includes: When the first image and / or the second image is detected, the visual detection device notifies the driving device in the visual verification device to move the verification piece out of the verification position.

16. The visual inspection method according to any one of claims 11 to 15, wherein: The visual detection method also includes: When the system verification result is normal, the host computer notifies the station controller to start conveying materials; When the result of the systematic check is abnormal, the host computer notifies each detection device to perform a shutdown test and issues an alarm.

17. The visual inspection method according to claim 12, wherein: The visual detection method also includes: The host computer performs visual accuracy verification according to the first image to obtain a visual accuracy verification result; The host computer performs imaging effect verification according to the second image to obtain an imaging effect verification result; When the visual accuracy verification result and the imaging effect verification result are both normal verification results, the host computer performs visual algorithm verification according to the target image to obtain a visual algorithm verification result; The host computer determines the systematic verification result of the visual inspection device according to the visual accuracy verification result, the imaging effect verification result and the visual algorithm verification result.

18. The visual inspection method according to claim 17, wherein: The host computer performs visual accuracy verification according to the first image to obtain a visual accuracy verification result, including: The host computer determines a first difference between a measurement value corresponding to a structural parameter of the size verification part in the first image and a preset standard value; The host computer obtains a verification result of abnormal visual accuracy of the visual detection device when the first difference is greater than or equal to a first parameter threshold range; When the first difference is less than a first parameter threshold range, the host computer obtains a verification result indicating that the visual accuracy of the visual inspection device is normal.

19. The visual inspection method according to claim 18, wherein: Also includes: The host computer determines a first parameter threshold range according to a tolerance corresponding to the structural parameter.

20. The visual inspection method according to claim 18, wherein: The visual detection method also includes: The host computer detects the first image through a target vision algorithm to obtain a target standard value corresponding to the structural parameter; The host computer determines a second difference between a measurement value corresponding to a structural parameter of the size verification part in the first image and a target standard value; The host computer obtains a verification result of an abnormality of the visual algorithm of the visual detection device when the second difference is greater than or equal to a second parameter threshold range; When the second difference is less than a second parameter threshold range, the host computer obtains a verification result indicating that the visual algorithm of the visual detection device is normal.

21. The visual inspection method according to claim 17, wherein: The host computer performs imaging effect verification according to the second image to obtain an imaging effect verification result, including: The host computer determines a third difference between the grayscale value of the color card verification part in the second image and a preset grayscale value; The host computer determines, when the third difference is greater than or equal to a third parameter threshold range, a detection result that the imaging effect of the visual detection device is abnormal; When the third difference is less than a third parameter threshold range, the host computer determines a detection result that the imaging effect of the visual detection device is normal.

22. The visual inspection method according to any one of claims 11 to 21, wherein: The visual detection method also includes: The host computer detects the second image by using a target vision algorithm to obtain a target grayscale value; The host computer determines a fourth difference between the grayscale value of the color card verification part in the second image and the target grayscale value; The host computer obtains a verification result of an abnormality in the visual algorithm of the visual detection device when the fourth difference is greater than or equal to a fourth parameter threshold range; When the fourth difference is less than the fourth parameter threshold range, the host computer obtains a verification result that the visual algorithm of the visual detection device is normal.

23. The visual inspection method according to claim 17, wherein: The host computer performs visual algorithm verification according to the target image, and before obtaining the visual algorithm verification result, it also includes: The host computer obtains a first picture, a second picture, a third picture, and a fourth picture with calibrated parameter values, wherein the parameter value of the first picture is within a first range, the parameter value of the second picture is within a second range, the parameter value of the third picture is within a third range, and the parameter value of the third picture is within a fourth range, the first range is different from the second range, and the third range is different from the fourth range; The host computer establishes a picture verification library according to the first picture, the second picture, the third picture and the fourth picture.

24. The visual inspection method according to claim 23, wherein: The host computer establishes a picture verification library according to the first picture, the second picture, the third picture and the fourth picture, including: The host computer numbers the first picture, the second picture, the third picture and the fourth picture; The host computer establishes a picture verification library according to the numbered first picture, second picture, third picture and fourth picture.

25. The visual inspection method according to claim 23, wherein: Before the host computer obtains the first picture, the second picture, the third picture and the fourth picture with calibrated parameter values, the process further includes: The host computer uses a target visual detection algorithm to evaluate the sample image to obtain a first picture, a second picture, a third picture and a fourth picture with calibrated parameter values.

26. The visual inspection method according to claim 23, wherein: The host computer performs visual algorithm verification according to the target image to obtain a visual algorithm verification result, including: The host computer selects a target image from the image verification library; The host computer performs image recognition on the target image through the visual detection device to obtain a parameter value of the target image; The host computer compares the parameter value with the calibration parameter value corresponding to the target image; When the difference between the parameter value and the calibration parameter value corresponding to the target image exceeds the fifth parameter threshold range, the host computer determines that the visual algorithm is abnormal.

Images