A double disk appearance detection method and device

Abstract

The present invention provides a double-disc appearance detection method and device, which relates to the field of automated detection technology, wherein the method comprises: extracting feature points from a set of images of a first plastic product to obtain a feature point extraction result; determining detection surface position information, detection surface quantity information and detection feature set according to the feature point extraction result; inputting the detection surface position information, the detection surface quantity information and the detection feature set into a first detection scheme generation model to generate a first detection scheme; the double-disc control module operates based on the control parameters of the first detection scheme, performs appearance detection on the first plastic product, and obtains a first detection result; based on the first detection scheme, determines the detection step information, and according to the detection step information, gradually diverts the abnormal products detected in the first detection result. The technical problems of low efficiency and low detection precision and accuracy of appearance detection of plastic products in the prior art are solved.

Description

Technical Field

[0001] The present invention relates to the field of automated detection technology, and in particular to a double-disc appearance detection method and device. Background Art

[0002] In industrial production, plastic products processed on the assembly line need to be inspected for appearance before they flow into the next link for packing and shipment or before they are assembled with other plastic product parts. Observe whether the appearance of the plastic products has burrs, cracks, burns, bubbles, inconsistent gloss and other appearance defects. Plastic products with appearance defects should be remelted or harmlessly treated to avoid negative impacts of defective products on corporate credit.

[0003] The current method for inspecting the appearance of plastic products generally uses a CCD camera combined with manual flipping of the plastic product to obtain the product appearance image, and then determines whether the plastic product has appearance defects based on appearance image analysis.

[0004] Due to the diversity of the structural forms of plastic products themselves, the existing plastic product appearance inspection methods cannot effectively collect status and detect defects on each surface of the plastic products. The existing technology has technical problems such as low inspection efficiency and low inspection precision and accuracy for the appearance inspection of plastic products. Summary of the invention

[0005] The present application provides a double-disc appearance inspection method and device, which are used to solve the technical problems in the prior art of low inspection efficiency and low inspection precision and accuracy in appearance inspection of plastic products.

[0006] In view of the above problems, the present application provides a double-disc appearance detection method and device.

[0007] In a first aspect of the present application, a dual-disc appearance detection method is provided, the method comprising: based on an image acquisition module, collecting an image set of a first plastic product; performing feature point extraction on the image set to obtain a first feature point extraction result; determining detection surface position information, detection surface quantity information and a detection feature set based on the first feature point extraction result; obtaining a first detection scheme generation model; inputting the detection surface position information, the detection surface quantity information and the detection feature set into the first detection scheme generation model to generate a first detection scheme, the first detection scheme comprising a plurality of control parameters of the dual-disc control module; based on the first detection scheme, after the dual-disc control module receives the control parameters, performing appearance detection on the first plastic product to obtain a first detection result; based on the first detection scheme, determining detection step information, and gradually diverting abnormal products detected in the first detection result according to the detection step information.

[0008] According to a second aspect of the present application, a dual-disc appearance inspection device is provided, the device comprising: a first acquisition unit, for collecting an image set of a first plastic product based on an image acquisition module; a first execution unit, for extracting feature points from the image set to obtain a first feature point extraction result; a second execution unit, for determining detection surface position information, detection surface quantity information and a detection feature set based on the first feature point extraction result; a first generation unit, for obtaining a first detection scheme generation model; a second generation unit, for inputting the detection surface position information, the detection surface quantity information and the detection feature set into the first detection scheme generation model to generate a first detection scheme, wherein the first detection scheme includes a plurality of control parameters of the dual-disc control module; a first detection unit, for performing appearance inspection on the first plastic product based on the first detection scheme, after the dual-disc control module receives the control parameters, to obtain a first detection result; and a third execution unit, for determining detection step information based on the first detection scheme, and gradually diverting abnormal products detected in the first detection result according to the detection step information.

[0009] The third aspect of the present application provides a dual-disc appearance detection system, comprising: a processor, the processor is coupled to a memory, the memory is used to store a program, when the program is executed by the processor, the system executes the steps of the method described in the first aspect.

[0010] A fourth aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method described in the first aspect are implemented.

[0011] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0012] The method provided in the embodiment of the present application obtains a first feature point extraction result by extracting feature points from a set of images of a first plastic product; determines detection surface position information, detection surface quantity information and detection feature set based on the first feature point extraction result; inputs the detection surface position information, the detection surface quantity information and the detection feature set into a first detection scheme generation model to generate a first detection scheme; after the double-disc control module receives the control parameters of the first detection scheme, it performs appearance inspection on the first plastic product to obtain a first detection result; based on the first detection scheme, determines the detection step information, and gradually diverts the abnormal products detected in the first detection result according to the detection step information. The present application obtains a first feature point extraction result by extracting feature points from a set of images of a first plastic product; and specifically splits the first feature point extraction result into detection surface position information, detection surface quantity information and detection feature set to provide an interception site for subsequent acquisition of a specific image of the first plastic product. The detection surface position information, the detection surface quantity information and the detection feature set are input into the first detection scheme generation model to generate the first detection scheme, and a scheme for detecting a specific plastic product is obtained based on the model to avoid omissions in the acquisition of appearance images during the appearance detection of the plastic product. After receiving the control parameters of the first detection scheme, the dual-disc control module performs an appearance detection on the first plastic product to obtain a first detection result; based on the first detection scheme, the detection step information is determined, and according to the detection step information, the abnormal products detected in the first detection result are gradually diverted. The technical effect of automatically performing accurate and efficient appearance detection on plastic products and removing defective plastic products is achieved.

[0013] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A schematic diagram of a double-disc appearance inspection method provided in this application;

[0015] Figure 2 A schematic diagram of a process for obtaining a first detection result in a double-disc appearance detection method provided in the present application;

[0016] Figure 3 A schematic diagram of a process for obtaining detection results by performing multi-scale abnormal convolution feature recognition in a dual-disc appearance detection method provided in this application;

[0017] Figure 4 A schematic diagram of the structure of a double-disc appearance inspection device is provided for this application;

[0018] Figure 5 This is a schematic diagram of the structure of an exemplary electronic device of the present application.

[0019] Explanation of the figure marks: first obtaining unit 11, first executing unit 12, second executing unit 13, first generating unit 14, second generating unit 15, first detecting unit 16, third executing unit 17, electronic device 300, memory 301, processor 302, communication interface 303, bus architecture 304. DETAILED DESCRIPTION

[0020] The present application provides a double-disc appearance inspection method and device, which is used to solve the technical problems in the prior art of low inspection efficiency and low inspection precision and accuracy in appearance inspection of plastic products.

[0021] Application Overview

[0022] Due to the high plasticity and strong chemical stability of plastics, various plastic products such as parts or appliances are processed and widely used in various scenarios. At the same time, due to the diversity of plastic products, their appearances also vary greatly. When performing appearance inspection on plastic products using existing appearance inspection methods, it is easy to miss the inspection surface and inaccurately detect surface defects. There are technical problems in the prior art that the inspection efficiency of appearance inspection on plastic products is low and the inspection precision and accuracy are not high.

[0023] In response to the above technical problems, the overall idea of ​​the technical solution provided by this application is as follows:

[0024] The method provided in the present application obtains a first feature point extraction result by extracting feature points from a set of images of a first plastic product; determines detection surface position information, detection surface quantity information and a detection feature set based on the first feature point extraction result; inputs the detection surface position information, the detection surface quantity information and the detection feature set into a first detection scheme generation model to generate a first detection scheme; after receiving the control parameters of the first detection scheme, the dual-disc control module performs appearance inspection on the first plastic product to obtain a first detection result; based on the first detection scheme, determines the detection step information, and according to the detection step information, gradually diverts the abnormal products detected in the first detection result.

[0025] After introducing the basic principles of the present application, the technical solutions in the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described here. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0026] Embodiment 1

[0027] like Figure 1 As shown, the present application provides a dual-disc appearance detection method, wherein the method is applied to a dual-disc appearance detection device, the device includes a dual-disc control module and an image acquisition module, and the method includes:

[0028] S100: Based on the image acquisition module, an image set of a first plastic product is acquired;

[0029] Specifically, the image acquisition module is a high-quality camera device with high pixel count and no motion blur when capturing moving objects. The first plastic product is an industrial part or an independent device made of plastic as the main raw material.

[0030] Before the first plastic product enters the detection device for appearance detection, the image acquisition module acquires a large number of images of the appearance of the first plastic product to ensure that there is no visual blind spot when observing the first plastic product based on image information, and the large number of images constitute an image set of the first plastic product. The image set of the first plastic product can fully display the appearance state of the first plastic product.

[0031] S200: extracting feature points from the image set to obtain a first feature point extraction result;

[0032] Specifically, it should be understood that plastic products with different appearances are prone to different parts of appearance defects due to differences in production molds and processing techniques. Therefore, when performing appearance inspection of plastic products, image inspection with emphasis can be performed based on the appearance characteristics of the plastic product, and turning surfaces, gaps, connection points or detailed features can be inspected with emphasis. In this embodiment, the first feature point extraction result is all the specific locations that need to be inspected for the appearance inspection of the first plastic product. If the first plastic product meets all the detection information in the first feature extraction result when performing appearance defect inspection, the first plastic product has no appearance defects. Image feature point analysis and extraction are performed based on the image set that can reflect the appearance status of the first plastic product obtained by the image acquisition module to form an appearance inspection feature point set and obtain the first feature point extraction result.

[0033] Optionally, the first feature point extraction result may also be used to expand the appearance inspection feature points based on common defect forms and defect occurrence sites in the plastic product production process, thereby increasing the inspection precision when performing appearance inspection on the first plastic product.

[0034] S300: Determine detection surface position information, detection surface quantity information and detection feature set according to the first feature point extraction result;

[0035] Specifically, the detection surface position information is the relative visual position relationship between the outer surfaces of the first plastic product that need to be inspected for appearance when the first plastic product is placed under the image acquisition device at a certain angle for image information acquisition, including the surfaces that the user will always see under normal operation, the surfaces that the user will sometimes see, the surfaces that the user will rarely see, and the surfaces that the user will never see. The detection surface quantity information is the number of visible and invisible surfaces that need to be inspected for the appearance inspection of the first plastic product. The position and number of the detection surfaces determine to a certain extent how many rounds of inspection are required for the first plastic product to perform appearance defect inspection on all surfaces. The detection feature set is all the locations that need to be inspected on each detection surface when the first plastic product is inspected for appearance.

[0036] S400: Obtaining a first detection scheme generation model;

[0037] Specifically, the first detection scheme is a scheme that specifically includes detection items, detection steps and detection methods, and can comprehensively detect the appearance of the first plastic product. The first detection scheme generation model is a model that has multiple sets of plastic product attributes and multiple sets of plastic product appearance detection schemes corresponding to the plastic product attributes. Based on the first detection scheme model, a suitable detection scheme for appearance defect detection of the first plastic product can be obtained. Specifically, the first detection scheme includes parameter settings for the dual-disc control module and a diversion and sorting detection step.

[0038] S500: Inputting the detection surface position information, the detection surface quantity information and the detection feature set into a first detection scheme generation model to generate a first detection scheme, wherein the first detection scheme includes a plurality of groups of control parameters of the dual-disc control module;

[0039] Specifically, the first plastic product is subjected to appearance inspection and defective product sorting control on the assembly line through a crawler. The dual-disc control module is a dual-disc appearance inspection device, which diverts the first plastic product through the cooperation of a high-speed glass turntable and an assembly line to gradually remove defective products. The control parameter of the dual-disc control module is the rotation speed of the high-speed glass turntable. The detection surface position information, the detection surface quantity information and the detection feature set are input into the first detection scheme generation model to generate the first detection scheme. The first detection scheme is a specific detection step for performing appearance inspection on the first plastic product, including each detection step and the control parameters of the dual-disc control module in each detection step. In theory, the control parameters of the dual-disc control module in each step are different.

[0040] S600: Based on the first detection scheme, after receiving the control parameters, the dual-disc control module performs appearance detection on the first plastic product to obtain a first detection result;

[0041] Specifically, the first test result is a set of appearance defect results obtained after testing the test features of each test surface of the first plastic product, such as whether the test surface has burrs, cracks, burns, bubbles, inconsistent gloss, and other quality control defects. The dual-disc control module operates with the control parameters of the first test solution obtained in the first test solution generation model, performs appearance inspection on the first plastic product, and obtains the first test result.

[0042] S700: Based on the first detection scheme, determine detection step information, and gradually divert abnormal products in the first detection result according to the detection step information.

[0043] Specifically, the detection step information is how many specific detection steps are needed to perform defect detection on all detection surfaces of the first plastic product when detecting the first plastic product.

[0044] Exemplarily, the number of detection steps can be determined based on the number of detection surfaces of the first plastic product, and the detection time and detection process required in each detection step can be determined according to the first detection scheme and realized by changing the parameters of the dual-disc control module. If the number of detection surfaces is determined to be 5 according to the extraction result of the first feature in step S300, the first plastic product needs to be screened in five steps to divert plastic products with appearance defects. The parameters of the dual-disc control module of each detection step are controlled based on the first detection scheme, so that each detection step can gradually divert abnormal products detected in the first plastic product in an orderly manner.

[0045] The present application extracts feature points through an image set of a first plastic product to obtain a first feature point extraction result; and the first feature point extraction result is specifically split into detection surface position information, detection surface quantity information and detection feature set, to provide an interception site for the subsequent acquisition of a specific image of the first plastic product. The detection surface position information, the detection surface quantity information and the detection feature set are input into a first detection scheme generation model to generate a first detection scheme, and a scheme for detecting a specific plastic product is obtained based on the model to avoid omissions in the acquisition of appearance images during the appearance inspection of the plastic product. After receiving the control parameters of the first detection scheme, the dual-disc control module performs an appearance inspection on the first plastic product to obtain a first detection result; based on the first detection scheme, the detection step information is determined, and according to the detection step information, the abnormal products detected in the first detection result are gradually diverted. The technical effect of automatically performing accurate and efficient appearance inspection on plastic products and removing defective plastic products is achieved.

[0046] Furthermore, the method step S400 of obtaining the first detection scheme generation model provided in the present application further includes:

[0047] S410: Obtaining plastic product type information of the first enterprise;

[0048] S420: Collecting corresponding historical plastic product testing solution information and corresponding plastic product basic attribute information based on the plastic product type information;

[0049] S430: Performing a detection solution generation model training based on the basic attribute information of the plastic product and the historical plastic product detection solution information to form a detection solution generation model library;

[0050] S440: Obtain type information of the first plastic product, and based on the type information of the first plastic product, call the first detection scheme generation model from the detection scheme generation model library to obtain the first detection scheme generation model.

[0051] Specifically, it should be understood that in order to improve the market competitiveness of enterprises, plastic production enterprises often have multiple production lines to produce different types of plastic products. Different appearance inspection methods are often used for different types of plastic products to screen out plastic products with appearance defects. The difference between the appearance inspection methods of different types of plastic products lies in the different number of inspection surfaces, the specific feature sites that need to be inspected on each inspection surface are different, and the specific number of feature sites is also different.

[0052] In this embodiment, the first enterprise is a manufacturer that is commissioned to conduct the appearance inspection of the first plastic product, and the basic attribute information of the plastic product includes the detection surface position information, detection surface quantity information and detection feature set information of the plastic product. The type information of all plastic products currently produced and discontinued by the first enterprise is obtained; the basic attribute information of all plastic products currently produced and discontinued by the first enterprise and multiple historical inspection scheme information are obtained by using the plastic product type information as a search instruction.

[0053] Based on the basic property information of the plastic product and the historical plastic product testing solution information, a testing solution generation model is trained to form a testing solution generation model library;

[0054] The type information of the first plastic product is obtained, and based on the type information of the first plastic product, the first detection scheme generation model is called from the detection scheme generation model library to obtain the first detection scheme generation model.

[0055] Specifically, the appearance defect problem of the first plastic product needs to be determined based on an accurate detection scheme generation model. Since the training model is a model that can continuously perform self-training and learning according to different actual conditions, in simple terms, it is a mathematical model. Based on a large amount of training data, each set of training data in the training data includes: based on the plastic product type information, the corresponding historical plastic product detection scheme information and the corresponding plastic product basic attribute information and the identification information used to identify the historical plastic product detection scheme results are collected. The neural network model continuously corrects itself. When the output information of the neural network model reaches a predetermined accuracy rate / reaches a convergence state, the supervised learning process ends. By training the data of the neural network model and according to the characteristics of the training model that the data is more accurate after training, the first detection scheme generation model output by the detection scheme generation model is also more accurate, thereby achieving the technical effect of obtaining a first detection scheme generation model that can perform a comprehensive and complete appearance inspection on the first plastic product, so that the first plastic product can accurately and efficiently detect the appearance defect results after inputting the first detection scheme generation model.

[0056] Further, such as Figure 2 As shown, the first plastic product is subjected to appearance inspection to obtain a first inspection result. The method step S600 provided in the present application further includes:

[0057] S610: performing real-time image acquisition on the first plastic product based on the image acquisition module to obtain a first image set;

[0058] S620: Obtain an abnormal image convolution feature, and perform image recognition on the first image set based on the abnormal image convolution feature to obtain a first image recognition feature value;

[0059] S630: judging whether secondary recognition is required according to the first image recognition feature value;

[0060] S640: if necessary, obtaining supplementary image information of the image acquisition site, performing secondary image recognition based on the supplementary image information, and obtaining a second image recognition feature value;

[0061] S650: If not necessary, obtaining the first detection result based on the first image recognition feature value.

[0062] Specifically, the first image set is a plurality of image information obtained by performing image acquisition on each inspection surface of the first plastic product. The image abnormality convolution feature is based on the existing image convolution processing method, and the image to be convolved and the convolution kernel are mapped into two matrices: the image matrix to be convolved and the convolution kernel matrix, respectively, so that the convolution operation between the image to be convolved and the convolution kernel is converted into the multiplication operation of the image matrix to be convolved and the convolution kernel matrix, and the product matrix obtained by the matrix multiplication operation is reversely mapped, so as to realize feature extraction to obtain the image abnormality convolution feature result, and the image abnormality convolution feature result is compared with the real-time image of the first plastic product to determine whether the corresponding image feature of the current image has an abnormality (defect).

[0063] Based on the image acquisition module, real-time image acquisition is performed on the first plastic product to obtain a first image set; based on the image abnormality convolution feature, image recognition is performed on the first image set to obtain a first image recognition feature value. The first image recognition feature value reflects the image clarity and information capacity of the first image set and whether there are defects in the appearance of the plastic product. A threshold value for image recognition feature is preset. If the first image recognition feature value reaches the threshold value, it means that the image of the first plastic product obtained contains defect information, and the first detection result is obtained based on the first image recognition feature value. If the first image recognition feature value does not reach the threshold value, it means that the image initially obtained at the image acquisition site is not clear or has other shooting defects, and it is necessary to collect supplementary image information again at the image acquisition site through the image acquisition device, and perform secondary image recognition based on the supplementary image information to obtain a second image recognition feature value. The second image recognition feature value is compared with the image recognition feature threshold value, and the detection result is obtained or the detection result is obtained after image fusion according to the comparison result. Until the image information obtained at the image acquisition site can be used to compare with the image abnormality convolution feature to obtain the detection result. Then, the image detection result is obtained based on the image recognition feature value.

[0064] The image abnormality convolution feature is used to determine whether the collected image meets the appearance detection and recognition requirements, and after the collected image meets the appearance detection and recognition requirements, feature comparison is performed to determine the appearance defect recognition result reflected by the collected image. This prevents appearance images with no recognition value from being used for appearance defect recognition detection, thereby misjudging the detection result based on invalid images, resulting in incomplete appearance detection of the first plastic product and omissions. The technical effect of improving the accuracy of appearance detection and recognition of the first plastic product based on the image abnormality convolution feature is achieved.

[0065] Further, such as Figure 3 As shown, the method steps provided by the present application also include:

[0066] S810: Determine whether image fusion is required based on the second image recognition feature value;

[0067] S820: If necessary, no diversion is performed in the current detection step, and the detection of subsequent steps is continued;

[0068] S830: After obtaining the image acquisition results of all steps, perform image fusion to obtain first fused image information;

[0069] S840: After performing multi-scale abnormal convolution feature recognition on the first fused image information, a third image recognition feature value is obtained;

[0070] S850: Obtain the first detection result based on the third image recognition feature value.

[0071] Specifically, it should be understood that when the camera fails to capture the image required for detection or the captured target image does not meet the clarity required for image feature extraction, it is necessary to re-acquire the image to meet the requirements of image feature extraction.

[0072] In this embodiment, the image fusion is to extract the beneficial information in each channel to the maximum extent through image processing and computer technology from the image data of the same target collected by multiple source channels, and finally synthesize it into a high-quality image. Multi-scale abnormal convolution feature recognition is to collect image information of multiple sites for image convolution recognition to obtain a collection of image abnormal convolution recognition features.

[0073] Based on the second image recognition feature value, it is compared with the preset image recognition feature threshold of the recognition site. If the image feature recognition threshold is still not met, image fusion is required to improve the resolution of the original image; if the image detection result cannot be obtained in the current detection step for diversion, the plastic products with appearance defects are not diverted and removed in the current detection step. The first plastic product continues to be detected in subsequent steps, and after obtaining the image acquisition results of all steps, image fusion is performed to obtain the first fused image information; after performing multi-scale abnormal convolution feature recognition on the first fused image information, the third image recognition feature value is obtained; based on the third image recognition feature value, the first detection result is obtained.

[0074] This embodiment uses image fusion technology to improve the utilization rate of image information and enhance the resolution of original images, and combines multiple rounds of diversion to obtain multi-scale abnormal convolution features to identify, compare and detect image fusion results, thereby achieving the technical effect of accurately obtaining the first detection result of the first plastic product.

[0075] Furthermore, according to the detection step information, the abnormal products detected in the first detection result are gradually diverted, and the method step S700 provided in the present application also includes:

[0076] S710: Obtaining abnormal product diversion detection steps and abnormal product number information;

[0077] S720: Obtaining an abnormal number mean based on the abnormal product number information;

[0078] S730: Based on the abnormal product number information and the average number of abnormalities, weight allocation is performed on each detection step to obtain a first weight allocation result;

[0079] S740: Based on the first weight distribution result, adjusting the detection time of the abnormal product diversion detection step to obtain a first adjustment time parameter;

[0080] S750: Optimize the first detection scheme based on the first adjustment time parameter.

[0081] Specifically, it should be understood that within the first inspection scheme, not all appearance defects of the first plastic product obtained in each step are the same. Therefore, if the same inspection time is set for each inspection step, it may happen that in some inspection steps, there is ample time to obtain inspection results, and accurate inspection results are obtained to divert defective products; while in some inspection steps, there is tight time to obtain inspection results, and defective products may not be detected or may be detected but not diverted, resulting in production accidents in which defective products are mistakenly judged as intact products.

[0082] In this embodiment, based on the first detection scheme obtained in step S500, a specific abnormal product diversion detection step is obtained, and information on the number of abnormal products in each diversion detection step detected according to the first detection scheme in the historical information is obtained; based on the abnormal product number information, an average value of the number of abnormalities in each diversion detection step is obtained; based on the abnormal product number information and the average value of the abnormal number, weights are assigned to each detection step to obtain a first weight assignment result; based on the first weight assignment result, a weighted calculation is performed to adjust the detection time of the abnormal product diversion detection step to obtain a first adjustment time parameter; based on the first adjustment time parameter, the first detection scheme is optimized so that the detection steps with a larger number of abnormal detection results obtain more detection time, thereby correspondingly reducing the detection time of the detection steps with a smaller number of abnormal detection results.

[0083] This embodiment adjusts the time consumption of each diversion detection step according to the information on the number of abnormal products in each diversion detection step, so that the time allocation of each diversion detection step meets the diversion detection accuracy target within the step, and each diversion detection step can accurately identify abnormalities and divert defective products on the image information within the step, thereby achieving the technical effect of improving the accuracy and efficiency of appearance inspection of the first plastic product.

[0084] Furthermore, the method steps provided in this application also include:

[0085] S910: Obtaining design complexity information and production accuracy information of the first plastic product;

[0086] S920: Setting expected accuracy information based on the design complexity information and the product accuracy information;

[0087] S930: Review the detection accuracy based on the first detection result to obtain detection accuracy information;

[0088] S940: Perform deviation analysis on the detection accuracy information and the expected accuracy information to obtain a first deviation analysis coefficient.

[0089] Furthermore, the method steps provided in this application also include:

[0090] S950: If the first deviation analysis coefficient is higher than the fault tolerance coefficient, based on the first deviation analysis coefficient, the first detection solution generation model is optimized using a PSO algorithm to obtain a first optimized detection solution generation model;

[0091] S960: Generate a test plan for the same batch of products of the first plastic product according to the first optimized test plan generation model, and store the first optimized test plan generation model in the test plan generation model library.

[0092] Specifically, it should be understood that in industrial production, there is a certain tolerance for product defects, and minor defective products that do not affect the functional use of plastic products are retained for use. The expected accuracy is the proportion of residual unqualified products in the plastic product obtained after multi-step diversion inspection when the first plastic product is subjected to appearance inspection. The first deviation analysis coefficient is the degree of deviation of the test result of the first plastic product from the test result error acceptable to the enterprise. The PSO algorithm (population-based stochastic optimization technology algorithm) is an iterative algorithm that simulates the local optimal solution to find the global optimal solution.

[0093] Obtain the design complexity information and production accuracy information of the first plastic product; set the expected accuracy information based on the design complexity information and the production accuracy information; perform a test accuracy review based on the first test result to obtain the test accuracy information; perform a deviation analysis on the test accuracy information and the expected accuracy information to obtain a first deviation analysis coefficient. If the first deviation analysis coefficient is higher than the fault tolerance coefficient, then based on the first deviation analysis coefficient, use the PSO algorithm to optimize the first test solution generation model to obtain a first optimized test solution generation model; generate a test solution for the same batch of products of the first plastic product based on the first optimized test solution generation model, and store the first optimized test solution generation model in the test solution generation model library.

[0094] By comparing the actual appearance inspection error rate of the first plastic product with the preset error rate, it is determined whether the current appearance inspection scheme is capable of screening out defective plastic products to the maximum extent. Based on this, the original appearance inspection scheme is optimized through an optimization algorithm, thereby achieving the technical effect of accurate, efficient and high-precision appearance defect identification and detection of the first plastic product.

[0095] Embodiment 2

[0096] Based on the same inventive concept as the double-disc appearance detection method in the aforementioned embodiment, Figure 4 As shown, the present application provides a double-disc appearance detection device, wherein the device comprises:

[0097] A first obtaining unit 11 is used to collect an image set of a first plastic product based on an image collection module;

[0098] A first execution unit 12 is used to extract feature points from the image set to obtain a first feature point extraction result;

[0099] A second execution unit 13 is used to determine detection surface position information, detection surface quantity information and detection feature set according to the first feature point extraction result;

[0100] A first generating unit 14 is used to obtain a first detection scheme generating model;

[0101] A second generating unit 15 is used to input the detection surface position information, the detection surface quantity information and the detection feature set into a first detection scheme generating model to generate a first detection scheme, wherein the first detection scheme includes a plurality of groups of control parameters of the dual-disc control module;

[0102] A first detection unit 16, configured to perform appearance detection on the first plastic product based on the first detection scheme and obtain a first detection result after the dual-disc control module receives the control parameters;

[0103] The third execution unit 17 is used to determine detection step information based on the first detection scheme, and gradually divert the abnormal products in the first detection result according to the detection step information.

[0104] Furthermore, the device also includes:

[0105] A second obtaining unit is used to obtain the type information of plastic products of the first enterprise;

[0106] A first collecting unit, configured to collect corresponding historical plastic product testing solution information and corresponding plastic product basic attribute information based on the plastic product type information;

[0107] A first construction unit is used to perform a detection scheme generation model training based on the basic attribute information of the plastic product and the historical plastic product detection scheme information to form a detection scheme generation model library;

[0108] The third obtaining unit is used to obtain type information of the first plastic product, and based on the type information of the first plastic product, call the first detection scheme generation model from the detection scheme generation model library to obtain the first detection scheme generation model.

[0109] Furthermore, the device also includes:

[0110] A second acquisition unit, configured to acquire a real-time image of the first plastic product based on the image acquisition module to obtain a first image set;

[0111] A first recognition unit, configured to obtain an abnormal image convolution feature, and perform image recognition on the first image set based on the abnormal image convolution feature to obtain a first image recognition feature value;

[0112] A first judging unit, configured to judge whether secondary recognition is required according to the first image recognition feature value;

[0113] A second judgment unit is used to obtain supplementary image information of the image acquisition site if necessary, perform secondary image recognition based on the supplementary image information, and obtain a second image recognition feature value;

[0114] The third judgment unit is used to obtain the first detection result based on the first image recognition feature value if it is not necessary.

[0115] Furthermore, the device also includes:

[0116] A fourth judgment unit, used for judging whether image fusion is required based on the second image recognition feature value;

[0117] A fifth judgment unit, configured to, if necessary, not perform diversion in the current detection step and continue to perform detection in subsequent steps;

[0118] A fourth execution unit, configured to perform image fusion after obtaining the image acquisition results of all steps, to obtain first fused image information;

[0119] A second recognition unit, configured to obtain a third image recognition feature value after performing multi-scale abnormal convolution feature recognition on the first fused image information;

[0120] The fourth obtaining unit is used to obtain the first detection result based on the third image recognition feature value.

[0121] Furthermore, the device also includes:

[0122] A fifth obtaining unit, used to obtain abnormal product diversion detection steps and abnormal product number information;

[0123] A sixth obtaining unit, configured to obtain an abnormal number mean based on the abnormal product number information;

[0124] A fifth execution unit, configured to perform weight allocation on each detection step based on the abnormal product number information and the average value of the abnormal number, to obtain a first weight allocation result;

[0125] A sixth execution unit, configured to adjust the detection time of the abnormal product diversion detection step based on the first weight distribution result to obtain a first adjustment time parameter;

[0126] A seventh execution unit is used to optimize the first detection scheme based on the first adjustment time parameter.

[0127] Furthermore, the device also includes:

[0128] a seventh obtaining unit, configured to obtain design complexity information and production accuracy information of the first plastic product;

[0129] an eighth execution unit, configured to set expected accuracy information based on the design complexity information and the product accuracy information;

[0130] an eighth obtaining unit, configured to perform a detection accuracy review based on the first detection result to obtain detection accuracy information;

[0131] The ninth execution unit is used to perform a deviation analysis on the detection accuracy information and the expected accuracy information to obtain a first deviation analysis coefficient.

[0132] Furthermore, the device also includes:

[0133] a tenth execution unit, configured to optimize the first detection scheme generation model by using a PSO algorithm based on the first deviation analysis coefficient to obtain a first optimized detection scheme generation model if the first deviation analysis coefficient is higher than the fault tolerance coefficient;

[0134] The third generation unit is used to generate a model based on the first optimized detection scheme, generate a detection scheme for the same batch of products of the first plastic product, and store the first optimized detection scheme generation model in the detection scheme generation model library.

[0135] Embodiment 3

[0136] Based on the same inventive concept as the double-disc appearance detection method in the aforementioned embodiment, the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method in embodiment one is implemented.

[0137] Exemplary Electronic Devices

[0138] Reference below Figure 5 To describe the electronic device of the present application.

[0139] Based on the same inventive concept as a dual-disc appearance detection method in the aforementioned embodiment, the present application also provides a dual-disc appearance detection system, including: a processor, the processor is coupled to a memory, the memory is used to store a program, when the program is executed by the processor, the system executes the steps of the method described in embodiment one.

[0140] The electronic device 300 includes: a processor 302, a communication interface 303, and a memory 301. Optionally, the electronic device 300 may also include a bus architecture 304. The communication interface 303, the processor 302, and the memory 301 may be interconnected via the bus architecture 304; the bus architecture 304 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus architecture 304 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0141] The processor 302 may be a CPU, a microprocessor, an ASIC, or one or more integrated circuits for controlling the execution of the program of the present application.

[0142] The communication interface 303 uses any transceiver-like device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), wired access networks, etc.

[0143] The memory 301 may be a ROM or other type of static storage device that can store static information and instructions, a RAM or other type of dynamic storage device that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, an optical disc storage (including a compressed optical disc, a laser disc, an optical disc, a digital versatile disc, a Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory may exist independently and be connected to the processor via a bus architecture 304. The memory may also be integrated with the processor.

[0144] The memory 301 is used to store computer-executable instructions for executing the solution of the present application, and the execution is controlled by the processor 302. The processor 302 is used to execute the computer-executable instructions stored in the memory 301, thereby realizing a double-disc appearance detection method provided in the above embodiment of the present application.

[0145] Those skilled in the art will understand that the various digital numbers such as the first and second involved in this application are only for the convenience of description, and are not used to limit the scope of this application, nor do they indicate a sequence of precedence. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects associated with each other are in an "or" relationship. "At least one" refers to one or more. At least two refers to two or more. "At least one", "any one" or similar expressions refer to any combination of these items, including any combination of single items (individuals) or plural items (individuals). For example, at least one item (individual, kind) of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, c can be single or multiple.

[0146] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)), etc.

[0147] The various illustrative logic units and circuits described in this application can be implemented or operated by a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination of the above. The general-purpose processor can be a microprocessor, and optionally, the general-purpose processor can also be any conventional processor, controller, microcontroller or state machine. The processor can also be implemented by a combination of computing devices, such as a digital signal processor and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a digital signal processor core, or any other similar configuration.

[0148] The steps of the method or algorithm described in this application can be directly embedded in hardware, software units executed by a processor, or a combination of the two. The software unit can be stored in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, register, hard disk, removable disk, CD-ROM or any other form of storage medium in the art. Exemplarily, the storage medium can be connected to the processor so that the processor can read information from the storage medium and write information to the storage medium. Optionally, the storage medium can also be integrated into the processor. The processor and the storage medium can be arranged in an ASIC, and the ASIC can be arranged in a terminal. Optionally, the processor and the storage medium can also be arranged in different components in the terminal. These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are performed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0149] Although the present application has been described in conjunction with specific features and embodiments thereof, it is obvious that various modifications and combinations may be made thereto without departing from the spirit and scope of the present application. Accordingly, this specification and the accompanying drawings are merely exemplary illustrations of the present application and are deemed to have covered any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A double disk appearance detection method, characterized in that: The method is applied to a dual-disc appearance detection device, the device comprising a dual-disc control module and an image acquisition module, and the method comprises: Based on the image acquisition module, an image set of a first plastic product is acquired; Extracting feature points from the image set to obtain a first feature point extraction result; Determine detection surface position information, detection surface quantity information and detection feature set according to the first feature point extraction result; Obtaining a first detection scheme generation model, specifically including: Obtain information on the type of plastic products of the first enterprise; Based on the plastic product type information, corresponding historical plastic product testing solution information and corresponding plastic product basic attribute information are collected; Based on the basic property information of the plastic product and the historical plastic product testing solution information, a testing solution generation model is trained to form a testing solution generation model library; obtaining type information of the first plastic product, and calling the first detection scheme generation model from the detection scheme generation model library based on the type information of the first plastic product to obtain the first detection scheme generation model; Inputting the detection surface position information, the detection surface quantity information and the detection feature set into a first detection scheme generation model to generate a first detection scheme, wherein the first detection scheme includes a plurality of groups of control parameters of the dual-disc control module; Based on the first detection scheme, after receiving the control parameters, the dual-disc control module performs appearance detection on the first plastic product to obtain a first detection result, which specifically includes: Based on the image acquisition module, real-time image acquisition is performed on the first plastic product to obtain a first image set; Obtaining an abnormal image convolution feature, and performing image recognition on the first image set based on the abnormal image convolution feature to obtain a first image recognition feature value; Determining whether secondary recognition is required according to the first image recognition feature value; If necessary, obtaining supplementary image information of the image acquisition site, performing secondary image recognition based on the supplementary image information, and obtaining a second image recognition feature value; If not required, obtaining the first detection result based on the first image recognition feature value; The obtaining of the second image recognition feature value specifically includes: Based on the second image recognition feature value, determining whether image fusion is required; If necessary, no diversion is performed in the current detection step, and the subsequent step detection is continued; After obtaining the image acquisition results of all steps, image fusion is performed to obtain first fused image information; After performing multi-scale abnormal convolution feature recognition on the first fused image information, a third image recognition feature value is obtained; Based on the third image recognition feature value, obtaining the first detection result; Determining detection step information based on the first detection scheme, and gradually diverting abnormal products in the first detection result according to the detection step information, specifically including: Obtain abnormal product diversion detection steps and abnormal product number information; Obtaining an abnormal number mean based on the abnormal product number information; Based on the abnormal product number information and the average number of abnormalities, weight allocation is performed on each detection step to obtain a first weight allocation result; Based on the first weight distribution result, adjusting the detection time of the abnormal product diversion detection step to obtain a first adjustment time parameter; Optimizing the first detection scheme based on the first adjustment time parameter; Obtaining design complexity information and production accuracy information of the first plastic product; Based on the design complexity information and the product accuracy information, setting expected accuracy information; Performing a detection accuracy review based on the first detection result to obtain detection accuracy information; Performing a deviation analysis on the detection accuracy information and the expected accuracy information to obtain a first deviation analysis coefficient; If the first deviation analysis coefficient is higher than the fault tolerance coefficient, based on the first deviation analysis coefficient, the first detection scheme generation model is optimized using a PSO algorithm to obtain a first optimized detection scheme generation model; According to the first optimized detection scheme generation model, a detection scheme is generated for the first plastic product and the same batch of products, and the first optimized detection scheme generation model is stored in the detection scheme generation model library.

2. A dual-disc appearance detection device for implementing the dual-disc appearance detection method according to claim 1, characterized in that: The device comprises: A first obtaining unit, configured to collect an image set of a first plastic product based on an image collection module; A first execution unit, configured to extract feature points from the image set to obtain a first feature point extraction result; A second execution unit, used to determine detection surface position information, detection surface quantity information and a detection feature set according to the first feature point extraction result; A first generating unit, used to obtain a first detection scheme generating model; A second generating unit, configured to input the detection surface position information, the detection surface quantity information and the detection feature set into a first detection scheme generating model to generate a first detection scheme, wherein the first detection scheme includes a plurality of groups of control parameters of a dual-disc control module; A first detection unit, configured to perform appearance detection on the first plastic product based on the first detection scheme and obtain a first detection result after the double-disc control module receives the control parameter; The third execution unit is used to determine the detection step information based on the first detection scheme, and gradually divert the abnormal products detected in the first detection result according to the detection step information.

3. A dual-disc appearance inspection system, characterized in that: The system comprises: a processor coupled to a memory, wherein the memory is used to store a program, and when the program is executed by the processor, the system is enabled to perform the steps of the method as claimed in claim 1.

4. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, which implements the steps of the method according to claim 1 when executed by a processor.

Images