Molybdenum plate detection method and system

Abstract

The invention relates to the technical field of material detection, and discloses a molybdenum plate detection method and system, and the method comprises the steps: obtaining a to-be-detected molybdenum plate, and collecting a visible light image and an ultrasonic image of the to-be-detected molybdenum plate through a pre-constructed image equipment cluster; according to a preset data processing strategy, carrying out noise reduction processing on the visible light image and the ultrasonic image to obtain noise reduction image data, carrying out molybdenum plate defect identification operation on the noise reduction image data to obtain a surface defect area, an internal defect area, a surface defect type and an internal defect type, and carrying out visual representation to obtain a molybdenum plate defect distribution image; and performing feature extraction and performance index score prediction operation on the molybdenum plate defect distribution image by using a molybdenum plate availability quality detection model to obtain a molybdenum plate performance prediction sequence, and comparing the molybdenum plate performance prediction sequence with an application scene performance requirement sequence to obtain a quality detection result. The defect detection accuracy of the molybdenum plate can be improved, and the utilization rate of the molybdenum plate material is increased.

Description

Technical Field

[0001] The present invention relates to the technical field of material detection, and in particular to a method and system for detecting a molybdenum plate. Background Art

[0002] Molybdenum plate is a thin plate made of molybdenum (Mo) material. Molybdenum is a metal with high melting point, high strength and excellent conductivity. It was originally widely used in aerospace and chemical industries. Due to the continuous improvement of technology, molybdenum plates are now rapidly expanding from aerospace and military industries to the civilian industrial field. They are often used to manufacture vacuum devices, thermocouples, discharge tubes, heating elements, etc.

[0003] In the field of molybdenum plate inspection, such as aerospace, nuclear energy, electronic manufacturing and high-temperature furnaces, there are usually very strict requirements on the quality and performance of molybdenum plate materials, and defective molybdenum plates will be reworked. However, this has led to other application industries also reworking defective molybdenum plates without knowing whether the defective molybdenum plates can be put into use. The pursuit of perfection in molybdenum plates not only wastes material resources, but also increases production cycles and costs. Summary of the invention

[0004] The present invention provides a method for detecting a molybdenum plate, the main purpose of which is to improve the accuracy of quality detection of the molybdenum plate.

[0005] To achieve the above object, the present invention provides a method for detecting a molybdenum plate, comprising: Acquire a molybdenum plate to be inspected, and use a pre-built image device cluster to collect a visible light image and an ultrasonic image of the molybdenum plate to be inspected; According to a preset data processing strategy, the visible light image and the ultrasonic image are subjected to denoising to obtain denoised image data, a molybdenum plate defect recognition operation is performed on the denoised image data to obtain a surface defect area and an internal defect area, respectively, and a defect type marking operation is performed on the surface defect area and the internal defect area to obtain a surface defect type and an internal defect type, respectively, wherein the denoised image data includes a visible light denoised image and an ultrasonic denoised image; Using a pre-built augmented reality algorithm, the molybdenum plate defects are imaged according to the surface defect area, the surface defect type, the internal defect type and the internal defect area to obtain a molybdenum plate defect distribution image; Using a pre-trained molybdenum plate usability quality detection model to perform a feature extraction operation on the molybdenum plate defect distribution image to obtain a molybdenum plate defect distribution feature, and performing a performance prediction on the molybdenum plate to be detected according to the molybdenum plate defect distribution feature to obtain a molybdenum plate performance prediction sequence; Obtain a standard performance sequence of molybdenum plates, and obtain a performance requirement sequence of application scenarios of the molybdenum plates to be tested, compare the standard performance sequence of molybdenum plates with the predicted performance sequence of molybdenum plates, obtain a performance defect sequence of molybdenum plates, and obtain a quality inspection result of the molybdenum plates to be tested based on a judgment result of whether the performance defect sequence of the molybdenum plates meets the performance requirement sequence of application scenarios.

[0006] Optionally, according to a preset data processing strategy, the visible light image and the ultrasonic image are subjected to denoising to obtain denoised image data, a molybdenum plate defect recognition operation is performed on the denoised image data to obtain a surface defect area and an internal defect area, respectively, and a defect type marking operation is performed on the surface defect area and the internal defect area to obtain a surface defect type and an internal defect type, respectively, including: According to a preset data processing strategy, the visible light image is subjected to Gaussian filtering to obtain a visible light denoised image, and the ultrasonic image is subjected to filtering using a pre-constructed low-pass filter to obtain an ultrasonic denoised image; Defect recognition is performed on the visible light denoised image according to a pre-built Sobel operator to obtain surface defects of the molybdenum plate; Using a pre-built image segmentation algorithm to perform defect type identification and cutting operations on the surface defects of the molybdenum plate to obtain surface defect areas and surface defect types; Performing a waveform feature extraction operation based on a time domain signal on the ultrasonic image to obtain a time domain analysis result, and using a pre-built Fourier transform algorithm to transform the ultrasonic image into an ultrasonic frequency domain signal, and performing a frequency component analysis on the ultrasonic frequency domain signal to obtain a frequency domain analysis result; Defect identification is performed on the time domain analysis results and the frequency domain analysis results to obtain internal defect types and internal defect areas.

[0007] Optionally, the use of a pre-built augmented reality algorithm to perform molybdenum plate defect imaging according to the surface defect area, surface defect type, internal defect type and internal defect area to obtain a molybdenum plate defect distribution image includes: Using a pre-built augmented reality platform, constructing virtual objects for the surface defect area, surface defect type, internal defect type, and internal defect area to obtain a virtual defect object set, wherein the defect types of the virtual defect objects in the virtual defect object set are configured by different colors; Acquire a surrounding image of a pre-built augmented reality device, perform a molybdenum plate positioning operation based on the molybdenum plate to be detected on the surrounding image, and obtain a posture of the molybdenum plate; The molybdenum plate plane of the molybdenum plate posture is identified by using a pre-built plane detection algorithm, and the position and posture of the virtual defect object set is located according to the molybdenum plate plane to obtain a molybdenum plate defect distribution image.

[0008] Optionally, the pre-trained molybdenum plate usability quality detection model is used to perform a feature extraction operation on the molybdenum plate defect distribution image to obtain the molybdenum plate defect distribution characteristics, and the performance of the molybdenum plate to be detected is predicted according to the molybdenum plate defect distribution characteristics to obtain a molybdenum plate performance prediction sequence, including: Using a pre-trained molybdenum plate usability quality detection model, a convolution operation is performed on the molybdenum plate defect distribution image to obtain a convolution matrix set; Performing a maximum pooling operation on the convolution matrix set to obtain a pooling feature set; Flattening the pooled feature set to obtain a defect distribution feature of the molybdenum plate, wherein the defect distribution feature of the molybdenum plate includes a defect area, a defect number, and a defect type; The defect distribution characteristics of the molybdenum plate are predicted based on a preset performance index set to obtain a molybdenum plate performance prediction sequence.

[0009] Optionally, before performing a prediction operation on the defect distribution characteristics of the molybdenum plate based on a preset performance indicator set, the method further includes: Obtain a preset performance index set including tensile strength, fatigue strength and electrical conductivity, wherein the tensile strength is expressed as: , In the formula, represents the predicted tensile strength, Indicates standard tensile strength, It represents the influence coefficient of defect area on tensile strength. It represents the influence coefficient of the number of defects on the tensile strength, Indicates the influence coefficient of defect type on tensile strength, represents the defect area, represents the number of defects, Indicator of the impact of defect type; Wherein, the fatigue strength is expressed as: , In the formula, represents the predicted fatigue strength, represents the standard fatigue strength, represents the mean stress, represents the influence coefficient of defect area on fatigue strength, represents the influence coefficient of the number of defects on fatigue strength, The influence coefficient of defect type on fatigue strength; Wherein, the electrical conductivity is expressed as: , In the formula, represents the predicted conductivity, represents the standard conductivity, represents the influence coefficient of defect area on conductivity, represents the influence coefficient of the number of defects on the conductivity, represents the influence coefficient of defect type on conductivity, An exponential factor representing the effect of defects on conductivity.

[0010] Optionally, before performing a feature extraction operation on the molybdenum plate defect distribution image using a pre-trained molybdenum plate usability quality detection model, the method further includes: Obtain a molybdenum plate usability quality detection model, and obtain a detection evaluation sample set, and allocate the detection evaluation sample set into a training set, a verification set, and a test set according to a preset allocation ratio, wherein the detection evaluation sample set includes: a sample molybdenum plate image and a true label sequence corresponding to the sample molybdenum plate image; According to the pre-configured training hyperparameters, a target sample is extracted from the training set in turn, and the molybdenum plate availability quality detection model is used to perform forward network prediction on the target sample to obtain a prediction performance index score sequence; Using a pre-built cross entropy loss algorithm, a loss value sequence between the true label sequence of the target sample and the prediction performance indicator score sequence is calculated; Minimize the loss value sequence to obtain a network model parameter sequence, and perform a reverse update of the parameters of the molybdenum plate availability quality detection model according to the network model parameter sequence to obtain an updated molybdenum plate availability quality detection model; Determine whether all the test evaluation samples in the training set participate in the training; When not all the detection evaluation samples in the training set participate in the training, return to the above step of extracting one target sample from the training set in sequence; When all the detection and evaluation samples in the training set participate in the training, the updated molybdenum plate availability quality detection model is evaluated using the verification set to obtain a verification and evaluation result, and according to the verification and evaluation result, the training hyperparameter is updated to obtain an updated hyperparameter; Using the test set to test the updated molybdenum plate availability quality detection model, to obtain a test accuracy rate; Determine whether the test accuracy is greater than a preset pass threshold; When the test accuracy is less than or equal to the qualified threshold, the hyperparameter is updated to the training hyperparameter, and the process returns to the step of allocating the detection evaluation sample set into a training set, a validation set, and a test set according to a preset allocation ratio; When the test accuracy is greater than the qualified threshold, the training process is stopped to obtain a trained molybdenum plate usability quality detection model.

[0011] Optionally, the molybdenum plate performance defect sequence is obtained by comparing the molybdenum plate standard performance sequence with the molybdenum plate performance prediction sequence, including: Performing difference calculation on the molybdenum plate standard performance sequence and the molybdenum plate performance prediction sequence to obtain a difference sequence set; Obtaining all performance labels whose difference scores are not zero in the difference sequence set to obtain a performance defect label set; The performance prediction score of each performance defect label in the performance defect label set in the molybdenum plate performance prediction sequence is extracted to obtain the molybdenum plate performance defect sequence.

[0012] Optionally, before obtaining the molybdenum plate standard performance sequence, the method further includes: Comparing the pre-built minimum defect threshold sequence with the molybdenum plate performance prediction sequence to obtain a comparison result; According to the comparison result, it is determined whether there is a performance tag in the molybdenum plate performance prediction sequence whose performance prediction score is less than the lowest defect threshold in the lowest defect threshold sequence; When there is a performance label in the molybdenum plate performance prediction sequence with a performance prediction score less than the lowest defect threshold in the lowest defect threshold sequence, the molybdenum plate to be tested is determined to be an unqualified product; When there is no performance tag in the molybdenum plate performance prediction sequence whose performance prediction score is less than the lowest defect threshold in the lowest defect threshold sequence, an operation of obtaining a standard performance sequence of the molybdenum plate is performed.

[0013] Optionally, after obtaining the quality test result of the molybdenum plate to be tested, the method further includes: Determining whether the quality inspection result is unqualified; When the quality inspection result is unqualified, the molybdenum plate type of the molybdenum plate to be inspected is obtained, wherein the molybdenum plate type includes a customized part molybdenum plate and a general size molybdenum plate; Determine whether the molybdenum plate type is a universal size molybdenum plate; When the molybdenum plate type is the universal size molybdenum plate, a maximum circumscribed matrix interception operation is performed on the molybdenum plate to be tested according to the application scenario performance requirement sequence to obtain a set of qualified molybdenum plates.

[0014] To achieve the above object, the present invention also provides a molybdenum plate detection system, comprising: A molybdenum plate image acquisition module, used to acquire the molybdenum plate to be detected, and to use a pre-built image device cluster to acquire a visible light image and an ultrasonic image of the molybdenum plate to be detected; A defect image recognition module is used to perform noise reduction processing on the visible light image and the ultrasonic image according to a preset data processing strategy to obtain noise reduction image data, perform a molybdenum plate defect recognition operation on the noise reduction image data to obtain a surface defect area and an internal defect area respectively, and perform a defect type marking operation on the surface defect area and the internal defect area to obtain a surface defect type and an internal defect type respectively, wherein the noise reduction image data includes a visible light noise reduction image and an ultrasonic noise reduction image; A performance prediction module is used to use a pre-built augmented reality algorithm to perform molybdenum plate defect imaging according to the surface defect area, surface defect type, internal defect type and internal defect area to obtain a molybdenum plate defect distribution image, and use a pre-trained molybdenum plate usability quality detection model to perform a feature extraction operation on the molybdenum plate defect distribution image to obtain a molybdenum plate defect distribution feature, and perform performance prediction on the molybdenum plate to be detected according to the molybdenum plate defect distribution feature to obtain a molybdenum plate performance prediction sequence; The quality inspection and evaluation module is used to obtain the standard performance sequence of the molybdenum plate and the application scenario performance requirement sequence of the molybdenum plate to be inspected, compare the standard performance sequence of the molybdenum plate with the molybdenum plate performance prediction sequence, obtain the molybdenum plate performance defect sequence, and obtain the quality inspection result of the molybdenum plate to be inspected based on the judgment result of whether the molybdenum plate performance defect sequence meets the application scenario performance requirement sequence.

[0015] In order to solve the above problem, the present invention further provides an electronic device, the electronic device comprising: a memory storing at least one instruction; and The processor executes the instructions stored in the memory to implement the above-mentioned molybdenum plate detection method.

[0016] In order to solve the above problems, the present invention also provides a computer-readable storage medium, in which at least one instruction is stored. The at least one instruction is executed by a processor in an electronic device to implement the above-mentioned molybdenum plate detection method.

[0017] The present invention solves the problem described in the background technology. The present invention first identifies the surface defects of the molybdenum plate through visible light images, then identifies the internal defects of the template through ultrasonic images, and then configures the relative positions of the molybdenum plate defects in the molybdenum plate through augmented reality technology to obtain a molybdenum plate defect distribution image, and then uses a pre-trained molybdenum plate usability quality detection model to identify the predicted results of the molybdenum plate performance according to the molybdenum plate defect distribution image to obtain a molybdenum plate performance prediction sequence; finally, by querying the molybdenum plate standard performance sequence and the application scenario performance requirement sequence, it is determined whether it can be put into use. Therefore, the present invention can improve the quality detection accuracy of the molybdenum plate. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic flow chart of a molybdenum plate detection method provided by one embodiment of the present invention; Figure 2 A functional module diagram of a molybdenum plate detection system provided by one embodiment of the present invention; Figure 3 A schematic diagram of the structure of an electronic device for implementing the molybdenum plate detection method provided in one embodiment of the present invention.

[0019] Description of figure markings: 1. Electronic device; 10. Processor; 11. Memory; 12. Bus.

[0020] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0021] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0022] The embodiment of the present application provides a method for detecting a molybdenum plate. The execution subject of the molybdenum plate detection method includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided in the embodiment of the present application. In other words, the molybdenum plate detection method can be executed by software or hardware installed in a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.

[0023] Reference Figure 1 FIG. 1 is a flow chart of a method for detecting a molybdenum plate according to an embodiment of the present invention. In this embodiment, the method for detecting a molybdenum plate includes: S1. Obtain a molybdenum plate to be inspected, and use a pre-built image device cluster to collect a visible light image and an ultrasonic image of the molybdenum plate to be inspected.

[0024] In the embodiment of the present invention, the molybdenum plate can be a universal size molybdenum plate, or a customized part molybdenum plate constructed with a special mold or cut into a special shape. When the universal size molybdenum plate has defects, the qualified area can be retained by cutting. When the customized part molybdenum plate does not meet the performance requirements, it needs to be reprocessed and manufactured.

[0025] Furthermore, the image device cluster includes a high-definition camera and ultrasonic detection equipment.

[0026] Specifically, in the embodiment of the present invention, the molybdenum plate to be inspected is first photographed from multiple angles by a high-definition camera to obtain a visible light image, and then the internal structure of the molybdenum plate to be inspected is detected by an ultrasonic detection device to obtain an ultrasonic image.

[0027] S2. According to a preset data processing strategy, the visible light image and the ultrasonic image are subjected to denoising processing to obtain denoised image data, a molybdenum plate defect recognition operation is performed on the denoised image data to obtain a surface defect area and an internal defect area, respectively, and a defect type marking operation is performed on the surface defect area and the internal defect area to obtain a surface defect type and an internal defect type, respectively, wherein the denoised image data includes a visible light denoised image and an ultrasonic denoised image.

[0028] In an embodiment of the present invention, the data processing strategy refers to using different noise reduction methods for different images, such as using a Gaussian filtering algorithm for visible light images and using a low-pass filter for noise reduction for ultrasonic images. The Gaussian filtering algorithm is an image processing technology mainly used for denoising and blurring. The principle is: based on the Gaussian function, by performing a convolution operation on the image, the high-frequency noise in the image is suppressed, thereby achieving a smoothing effect. The low-pass filter (LPF) is a filter that allows low-frequency signals to pass through and blocks high-frequency signals, and is also used to process image noise reduction.

[0029] In detail, in an embodiment of the present invention, according to a preset data processing strategy, the visible light image and the ultrasonic image are subjected to denoising to obtain denoised image data, the denoised image data is subjected to a molybdenum plate defect recognition operation to obtain a surface defect area and an internal defect area, respectively, and the surface defect area and the internal defect area are subjected to a defect type marking operation to obtain a surface defect type and an internal defect type, respectively, including: According to a preset data processing strategy, the visible light image is subjected to Gaussian filtering to obtain a visible light denoised image, and the ultrasonic image is subjected to filtering using a pre-constructed low-pass filter to obtain an ultrasonic denoised image; Defect recognition is performed on the visible light denoised image according to a pre-built Sobel operator to obtain surface defects of the molybdenum plate; Using a pre-built image segmentation algorithm to perform defect type identification and cutting operations on the surface defects of the molybdenum plate to obtain surface defect areas and surface defect types; Performing a waveform feature extraction operation based on a time domain signal on the ultrasonic image to obtain a time domain analysis result, and using a pre-built Fourier transform algorithm to transform the ultrasonic image into an ultrasonic frequency domain signal, and performing a frequency component analysis on the ultrasonic frequency domain signal to obtain a frequency domain analysis result; Defect identification is performed on the time domain analysis results and the frequency domain analysis results to obtain internal defect types and internal defect areas.

[0030] The sobel operator is an edge detection algorithm, which is mainly used in image processing to detect edges by calculating the gradient of pixels in an image, especially in places where the intensity of the image varies greatly, and is usually used to extract edges in the horizontal and vertical directions. The embodiment of the present invention can identify special areas of the molybdenum plate, such as scratches and oxidized areas, through edge detection.

[0031] Furthermore, the image segmentation algorithm aims to divide the image into multiple parts or regions for analysis and processing. The present invention performs image segmentation based on the above-mentioned edge detection results (molybdenum plate surface defects) to obtain the surface defect area. In addition, the present invention can also identify the defect type based on the shape characteristics of the surface defect area, such as linear, flaky, protruding or concave features, to obtain the surface defect type.

[0032] Similarly, in the embodiment of the present invention, defect area identification and defect type identification are also performed on the ultrasonic image. The difference is that the ultrasonic image data type characteristics can be divided into time domain signals and frequency domain signals. In the embodiment of the present invention, the signal characteristics of the time domain signal and the frequency domain signal are extracted and identified successively, and finally the internal defect type and internal defect area are obtained. The waveform feature extraction operation of the time domain signal refers to extracting information such as peak value and envelope to obtain the time domain analysis result, while the frequency component analysis of the ultrasonic frequency domain signal refers to analyzing the energy of different frequency components to obtain the frequency domain analysis result.

[0033] The ultrasonic image itself is a time domain signal, and the present invention can convert it through the Fourier transform algorithm. The Fourier transform is expressed as: , In the formula, represents the ultrasonic frequency domain signal, represents the ultrasound image, Indicates frequency, Indicates time.

[0034] S3. Using a pre-built augmented reality algorithm, the molybdenum plate defects are imaged according to the surface defect area, surface defect type, internal defect type and internal defect area to obtain a molybdenum plate defect distribution image.

[0035] Among them, the augmented reality algorithm (AR for short) is a technology that superimposes virtual information (such as images, sounds, videos and other data) into the real world, thereby enhancing the user's perception and interaction with the real world.

[0036] The present invention configures the molybdenum plate defects onto the molybdenum plate to be inspected through augmented reality. On the one hand, it is convenient for inspectors to observe the physical molybdenum plate at any time to facilitate quality monitoring. On the other hand, a 3D-dimensional molybdenum plate defect distribution image can be provided to facilitate subsequent models to predict performance according to the defect distribution.

[0037] In detail, in an embodiment of the present invention, the use of a pre-built augmented reality algorithm to perform molybdenum plate defect imaging according to the surface defect area, surface defect type, internal defect type and internal defect area to obtain a molybdenum plate defect distribution image includes: Using a pre-built augmented reality platform, constructing virtual objects for the surface defect area, surface defect type, internal defect type, and internal defect area to obtain a virtual defect object set, wherein the defect types of the virtual defect objects in the virtual defect object set are configured by different colors; Acquire a surrounding image of a pre-built augmented reality device, perform a molybdenum plate positioning operation based on the molybdenum plate to be detected on the surrounding image, and obtain a posture of the molybdenum plate; The molybdenum plate plane of the molybdenum plate posture is identified by using a pre-built plane detection algorithm, and the position and posture of the virtual defect object set is located according to the molybdenum plate plane to obtain a molybdenum plate defect distribution image.

[0038] Specifically, in the embodiment of the present invention, the Unity augmented reality platform is first used, and then the required SDK and development environment are installed, and the development tools are configured, and then the surface defect area, surface defect type, internal defect type and internal defect area defect data are input, and the augmented reality platform will parse these data to generate corresponding virtual defect objects. The present invention can fill virtual defect objects of different defect types according to different colors or color depths. Among them, Unity can be used across platforms, and if iOS devices or other operating systems such as Android can use AR platforms such as ARKit or ARCore.

[0039] Furthermore, the augmented reality device refers to the user's AR glasses and other devices. The user can wear the augmented reality device to observe the defects of the molybdenum plate. In the embodiment of the present invention, the image observed by the user through the augmented reality device can be called the surrounding image, and then the molybdenum plate to be detected is automatically identified and located through the object recognition method of the neural network to obtain the posture of the molybdenum plate. Then, the augmented reality device can use the plane detection algorithm to observe the real posture of the molybdenum plate to be detected, and then according to a special point in the molybdenum plate as the coordinate base point, the position and posture of the molybdenum plate defects are adaptively located and configured to obtain the molybdenum plate defect distribution image. It is convenient for the inspection personnel to observe the defect distribution of one or more molybdenum plates to be inspected. Among them, the plane detection algorithm is a type of technology used to identify and extract plane structures in three-dimensional space, which will not be elaborated here.

[0040] S4. Use the pre-trained molybdenum plate availability quality detection model to perform feature extraction operation on the molybdenum plate defect distribution image to obtain the molybdenum plate defect distribution characteristics, and perform performance prediction on the molybdenum plate to be detected according to the molybdenum plate defect distribution characteristics to obtain a molybdenum plate performance prediction sequence.

[0041] The molybdenum plate usability quality detection model is a multi-task model based on a neural network, which can evaluate and predict some molybdenum plate properties according to the defect characteristics of the molybdenum plate. The properties to be evaluated in the embodiment of the present invention include tensile strength, fatigue strength and electrical conductivity. In addition, some performance indicators such as thermal conductivity and hardness can be added through later training.

[0042] In detail, in an embodiment of the present invention, the pre-trained molybdenum plate usability quality detection model is used to perform a feature extraction operation on the molybdenum plate defect distribution image to obtain the molybdenum plate defect distribution characteristics, and the performance of the molybdenum plate to be detected is predicted according to the molybdenum plate defect distribution characteristics to obtain a molybdenum plate performance prediction sequence, including: Using a pre-trained molybdenum plate usability quality detection model, a convolution operation is performed on the molybdenum plate defect distribution image to obtain a convolution matrix set; Performing a maximum pooling operation on the convolution matrix set to obtain a pooling feature set; Flattening the pooled feature set to obtain a defect distribution feature of the molybdenum plate, wherein the defect distribution feature of the molybdenum plate includes a defect area, a defect number, and a defect type; The defect distribution characteristics of the molybdenum plate are predicted based on a preset performance index set to obtain a molybdenum plate performance prediction sequence.

[0043] In an embodiment of the present invention, the convolution operation is calculated by a preset convolution kernel, each convolution kernel is used to extract a defect feature, and finally a convolution matrix set is obtained, and then the convolution matrix set is reduced in dimension by a maximum pooling operation to obtain a pooling feature set. Then the present invention further performs a dimensionality reduction operation using a flattening layer to obtain a one-dimensional molybdenum plate defect distribution feature. Finally, various performance indicators are calculated by a fully connected layer to obtain a molybdenum plate performance prediction sequence. Among them, the maximum pooling operation refers to a downsampling technology, which is mainly used to reduce the size of the feature map, thereby reducing the computational complexity and improving the model's anti-noise ability; the flattening operation is used to convert multidimensional data into one-dimensional data.

[0044] Furthermore, in an embodiment of the present invention, before the prediction operation of the molybdenum plate defect distribution characteristics based on a preset performance indicator set is performed, the method further includes: Obtain a preset performance index set including tensile strength, fatigue strength and electrical conductivity, wherein the tensile strength is expressed as: , In the formula, represents the predicted tensile strength, Indicates standard tensile strength, It represents the influence coefficient of defect area on tensile strength, It represents the influence coefficient of the number of defects on the tensile strength, Indicates the influence coefficient of defect type on tensile strength, represents the defect area, represents the number of defects, Indicator of the impact of defect type; Wherein, the fatigue strength is expressed as: , In the formula, represents the predicted fatigue strength, represents the standard fatigue strength, represents the mean stress, represents the influence coefficient of defect area on fatigue strength, represents the influence coefficient of the number of defects on fatigue strength, The influence coefficient of defect type on fatigue strength; Wherein, the electrical conductivity is expressed as: , In the formula, represents the predicted conductivity, represents the standard conductivity, represents the influence coefficient of defect area on conductivity, represents the influence coefficient of the number of defects on the conductivity, represents the influence coefficient of defect type on conductivity, An exponential factor representing the effect of defects on conductivity.

[0045] In the embodiment of the present invention, three indicators, namely defect area, defect number and defect type, are quantified, so that a linear function is constructed by training model parameters such as k, m, t, f, g and h, so as to predict the score of each performance indicator.

[0046] Furthermore, in an embodiment of the present invention, before performing a feature extraction operation on the molybdenum plate defect distribution image using the pre-trained molybdenum plate availability quality detection model, the method further includes: Obtain a molybdenum plate usability quality detection model, and obtain a detection evaluation sample set, and allocate the detection evaluation sample set into a training set, a verification set, and a test set according to a preset allocation ratio, wherein the detection evaluation sample set includes: a sample molybdenum plate image and a true label sequence corresponding to the sample molybdenum plate image; According to the pre-configured training hyperparameters, a target sample is extracted from the training set in turn, and the molybdenum plate availability quality detection model is used to perform forward network prediction on the target sample to obtain a prediction performance index score sequence; Using a pre-built cross entropy loss algorithm, a loss value sequence between the true label sequence of the target sample and the prediction performance indicator score sequence is calculated; Minimize the loss value sequence to obtain a network model parameter sequence, and perform a reverse update of the parameters of the molybdenum plate availability quality detection model according to the network model parameter sequence to obtain an updated molybdenum plate availability quality detection model; Determine whether all the test evaluation samples in the training set participate in the training; When not all the detection evaluation samples in the training set participate in the training, return to the above step of extracting one target sample from the training set in sequence; When all the detection and evaluation samples in the training set participate in the training, the updated molybdenum plate availability quality detection model is evaluated using the verification set to obtain a verification and evaluation result, and according to the verification and evaluation result, the training hyperparameter is updated to obtain an updated hyperparameter; Using the test set to test the updated molybdenum plate availability quality detection model, to obtain a test accuracy rate; Determine whether the test accuracy is greater than a preset pass threshold; When the test accuracy is less than or equal to the qualified threshold, the hyperparameter is updated to the training hyperparameter, and the process returns to the step of allocating the detection evaluation sample set into a training set, a validation set, and a test set according to a preset allocation ratio; When the test accuracy is greater than the qualified threshold, the training process is stopped to obtain a trained molybdenum plate usability quality detection model.

[0047] In an embodiment of the present invention, the detection and evaluation sample set includes: a sample molybdenum plate image and a real label sequence corresponding to the sample molybdenum plate image. Then, the detection and evaluation sample set is grouped according to a preset allocation ratio, such as a ratio of 8:1:1, to obtain a training set, a validation set, and a test set. Among them, the real label sequence refers to the real detection results of various performance indicators of the sample molybdenum plate, that is, [molybdenum plate 1: real tensile strength, real fatigue strength and real conductivity; molybdenum plate 2: ...; molybdenum plate n: ...].

[0048] Among them, the training set is used to train the model, the validation set is used to adjust the training hyperparameters, such as learning rate, regularization strength, etc. during the training process to improve the generalization ability of the model, and the test set is used to control the accuracy of the model after training.

[0049] The embodiment of the present invention performs forward network calculation on the model according to the training set to obtain a prediction performance index score sequence, wherein the prediction performance index score sequence includes predicted tensile strength, predicted fatigue strength and predicted conductivity.

[0050] Specifically, in an embodiment of the present invention, the model training direction is first controlled by a cross entropy loss algorithm and a gradient descent algorithm, wherein the result processes of the various indicators in the model are independent of each other, and therefore have multiple independent loss values, and finally a loss value sequence is obtained, wherein the loss value sequence includes [a first loss value corresponding to the tensile strength prediction, a second loss value corresponding to the fatigue strength prediction, and a third loss value corresponding to the conductivity prediction]. The present invention needs to minimize each loss value in the loss value sequence, obtain the network model parameters when each loss value is the smallest, wherein the network model parameters include parameters such as k, m, t, f, g, and h, and then perform a reverse update on the network model parameters to obtain an updated molybdenum plate availability quality detection model.

[0051] Then, after each training cycle (epoch), the current model is evaluated on the validation set to obtain the validation evaluation results, where the validation evaluation results include accuracy and recall, etc. Then, according to the validation evaluation results of the validation set, the hyperparameters (such as learning rate, regularization strength, etc.) are adjusted. If the model is overfitting, an early stopping strategy can be used.

[0052] Finally, the test set is imported into the molybdenum plate availability quality detection model to obtain the test structure, and then the proportion of correct predictions in the test results is checked to obtain the test accuracy. When the test accuracy is qualified, the training process is stopped to obtain the trained molybdenum plate availability quality detection model.

[0053] S5. Obtain a standard performance sequence of molybdenum plates, and obtain a performance requirement sequence of the application scenarios of the molybdenum plates to be tested, compare the standard performance sequence of the molybdenum plates with the predicted performance sequence of the molybdenum plates, obtain a performance defect sequence of the molybdenum plates, and obtain a quality inspection result of the molybdenum plates to be tested based on a judgment result of whether the performance defect sequence of the molybdenum plates meets the performance requirement sequence of the application scenarios.

[0054] The standard performance sequence of molybdenum plates refers to the highest standard performance indicators of molybdenum plates in the industry, including hardness, tensile strength, fatigue strength, etc. In the embodiment of the present invention, only tensile strength, fatigue strength and electrical conductivity are considered.

[0055] Furthermore, due to the difference in application scenarios, fields such as thermocouples and discharge tubes do not require too high hardness, but only require good conductivity and other indicators. Therefore, as long as the application scenario performance requirement sequence of the molybdenum plate to be tested is confirmed, some molybdenum plates with defects in performance indicators can still be used.

[0056] In detail, in the embodiment of the present invention, the molybdenum plate performance defect sequence is obtained by comparing the molybdenum plate standard performance sequence with the molybdenum plate performance prediction sequence, including: Performing difference calculation on the molybdenum plate standard performance sequence and the molybdenum plate performance prediction sequence to obtain a difference sequence set; Obtaining all performance labels whose difference scores are not zero in the difference sequence set to obtain a performance defect label set; The performance prediction score of each performance defect label in the performance defect label set in the molybdenum plate performance prediction sequence is extracted to obtain the molybdenum plate performance defect sequence.

[0057] In the embodiment of the present invention, the index alignment difference calculation is performed according to the order of the tensile strength prediction score, fatigue strength prediction score and conductivity prediction score in the molybdenum plate performance prediction sequence to obtain a difference sequence set. Then, the performance prediction scores of all performance labels with non-zero difference scores in the difference sequence set are sorted to obtain a molybdenum plate performance defect sequence. The molybdenum plate performance defect sequence contains defective indicators.

[0058] Specifically, in the embodiment of the present invention, the quality inspection result of the molybdenum plate to be inspected is obtained according to the judgment result of whether the molybdenum plate performance defect sequence meets the application scenario performance requirement sequence. Among them, when the molybdenum plate performance defect sequence cannot meet the application scenario performance requirement sequence, that is, a certain performance indicator in the molybdenum plate performance defect sequence appears in the application scenario performance requirement sequence, it means that the requirements of the molybdenum plate application scenario cannot be met, resulting in the molybdenum plate to be inspected being judged as unqualified. Otherwise, it is judged as qualified.

[0059] Furthermore, in an embodiment of the present invention, before obtaining the standard performance sequence of the molybdenum plate, the method further includes: Comparing the pre-built minimum defect threshold sequence with the molybdenum plate performance prediction sequence to obtain a comparison result; According to the comparison result, it is determined whether there is a performance tag in the molybdenum plate performance prediction sequence whose performance prediction score is less than the lowest defect threshold in the lowest defect threshold sequence; When there is a performance label in the molybdenum plate performance prediction sequence with a performance prediction score less than the lowest defect threshold in the lowest defect threshold sequence, the molybdenum plate to be tested is determined to be an unqualified product; When there is no performance tag in the molybdenum plate performance prediction sequence whose performance prediction score is less than the lowest defect threshold in the lowest defect threshold sequence, an operation of obtaining a standard performance sequence of the molybdenum plate is performed.

[0060] The minimum defect threshold sequence refers to the industry minimum requirement for molybdenum plates. Products below the minimum defect threshold sequence are judged as unqualified products and are not applicable to any application scenario. Therefore, in the embodiment of the present invention, the minimum defect threshold sequence can be obtained for preliminary judgment in sequence, and then the performance requirement sequence of the specific application scenario can be judged.

[0061] Further, in an embodiment of the present invention, after obtaining the quality inspection result of the molybdenum plate to be inspected, the method further includes: Determining whether the quality inspection result is unqualified; When the quality inspection result is unqualified, the molybdenum plate type of the molybdenum plate to be inspected is obtained, wherein the molybdenum plate type includes a customized part molybdenum plate and a general size molybdenum plate; Determine whether the molybdenum plate type is a universal size molybdenum plate; When the molybdenum plate type is the universal size molybdenum plate, a maximum circumscribed matrix interception operation is performed on the molybdenum plate to be tested according to the application scenario performance requirement sequence to obtain a set of qualified molybdenum plates.

[0062] According to the above, the types of molybdenum plates include customized parts molybdenum plates and universal size molybdenum plates. Universal size molybdenum plates can be cut, while customized parts molybdenum plates cannot be cut. Therefore, the qualified positions of universal size molybdenum plates can be cut to reduce the re-manufacturing ratio of molybdenum plate products and save the production cost of molybdenum plates.

[0063] The present invention solves the problem described in the background technology. The present invention first identifies the surface defects of the molybdenum plate through visible light images, then identifies the internal defects of the template through ultrasonic images, and then configures the relative positions of the molybdenum plate defects in the molybdenum plate through augmented reality technology to obtain a molybdenum plate defect distribution image, and then uses a pre-trained molybdenum plate usability quality detection model to identify the predicted results of the molybdenum plate performance according to the molybdenum plate defect distribution image to obtain a molybdenum plate performance prediction sequence; finally, by querying the molybdenum plate standard performance sequence and the application scenario performance requirement sequence, it is determined whether it can be put into use. Therefore, the present invention can improve the quality detection accuracy of the molybdenum plate.

[0064] like Figure 2 , which is a functional module diagram of a molybdenum plate detection system provided in one embodiment of the present invention.

[0065] The molybdenum plate detection system 100 of the present invention can be installed in an electronic device. According to the functions to be implemented, the molybdenum plate detection system 100 can include a molybdenum plate image acquisition module 101, a defect image recognition module 102, a performance prediction module 103 and a quality detection and evaluation module 104. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0066] The molybdenum plate image acquisition module 101 is used to acquire the molybdenum plate to be detected, and to use a pre-built image device cluster to acquire a visible light image and an ultrasonic image of the molybdenum plate to be detected; The defect image recognition module 102 is used to perform noise reduction processing on the visible light image and the ultrasonic image according to a preset data processing strategy to obtain noise reduction image data, perform a molybdenum plate defect recognition operation on the noise reduction image data to obtain a surface defect area and an internal defect area respectively, and perform a defect type marking operation on the surface defect area and the internal defect area to obtain a surface defect type and an internal defect type respectively, wherein the noise reduction image data includes a visible light noise reduction image and an ultrasonic noise reduction image; The performance prediction module 103 is used to use a pre-built augmented reality algorithm to perform molybdenum plate defect imaging according to the surface defect area, surface defect type, internal defect type and internal defect area to obtain a molybdenum plate defect distribution image, and use a pre-trained molybdenum plate availability quality detection model to perform a feature extraction operation on the molybdenum plate defect distribution image to obtain a molybdenum plate defect distribution feature, and perform performance prediction on the molybdenum plate to be detected according to the molybdenum plate defect distribution feature to obtain a molybdenum plate performance prediction sequence; The quality inspection and evaluation module 104 is used to obtain the standard performance sequence of the molybdenum plate and the application scenario performance requirement sequence of the molybdenum plate to be inspected, and compare the standard performance sequence of the molybdenum plate with the molybdenum plate performance prediction sequence to obtain the molybdenum plate performance defect sequence, and obtain the quality inspection result of the molybdenum plate to be inspected based on the judgment result of whether the molybdenum plate performance defect sequence meets the application scenario performance requirement sequence.

[0067] In detail, the modules in the molybdenum plate detection system 100 of the embodiment of the present invention are used in the same manner as described above. Figure 1 The detection method of the molybdenum plate described in the invention has the same technical means and can produce the same technical effect, so I will not go into details here.

[0068] like Figure 3 , which is a schematic diagram of the structure of an electronic device for implementing a molybdenum plate detection method provided by an embodiment of the present invention.

[0069] The electronic device 1 may include a processor 10, a memory 11 and a bus 12, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a molybdenum plate detection method program.

[0070] Wherein, the memory 11 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, mobile hard disk, multimedia card, card-type memory (for example: SD or DX memory, etc.), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as a mobile hard disk of the electronic device 1. In other embodiments, the memory 11 can also be an external storage device of the electronic device 1, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device 1. Further, the memory 11 also includes an internal storage unit of the electronic device 1 and an external storage device. The memory 11 can not only be used to store application software and various types of data installed in the electronic device 1, such as the code of the molybdenum plate detection method program, etc., but also can be used to temporarily store data that has been output or is to be output.

[0071] The processor 10 may be composed of an integrated circuit in some embodiments, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips, etc. The processor 10 is the control core (Control Unit) of the electronic device, and uses various interfaces and lines to connect various components of the entire electronic device, and executes or executes programs or modules (such as molybdenum plate detection method programs, etc.) stored in the memory 11, and calls data stored in the memory 11 to execute various functions of the electronic device 1 and process data.

[0072] The bus 12 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 may be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to realize connection and communication between the memory 11 and at least one processor 10, etc.

[0073] Figure 3 Only an electronic device with components is shown, and those skilled in the art will understand that Figure 3 The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0074] For example, although not shown, the electronic device 1 may also include a power source (such as a battery) for supplying power to various components. Preferably, the power source may be logically connected to the at least one processor 10 through a power management device, so that the power management device can realize functions such as charging management, discharging management, and power consumption management. The power source may also include one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 1 may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

[0075] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device 1 and other electronic devices.

[0076] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), or a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode) touch device. The display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device 1 and to display a visual user interface.

[0077] The molybdenum plate detection method program stored in the memory 11 in the electronic device 1 is a combination of multiple instructions. When running in the processor 10, it can achieve: Acquire a molybdenum plate to be inspected, and use a pre-built image device cluster to collect a visible light image and an ultrasonic image of the molybdenum plate to be inspected; According to a preset data processing strategy, the visible light image and the ultrasonic image are subjected to denoising to obtain denoised image data, a molybdenum plate defect recognition operation is performed on the denoised image data to obtain a surface defect area and an internal defect area, respectively, and a defect type marking operation is performed on the surface defect area and the internal defect area to obtain a surface defect type and an internal defect type, respectively, wherein the denoised image data includes a visible light denoised image and an ultrasonic denoised image; Using a pre-built augmented reality algorithm, the molybdenum plate defects are imaged according to the surface defect area, the surface defect type, the internal defect type and the internal defect area to obtain a molybdenum plate defect distribution image; Using a pre-trained molybdenum plate usability quality detection model to perform a feature extraction operation on the molybdenum plate defect distribution image to obtain a molybdenum plate defect distribution feature, and performing a performance prediction on the molybdenum plate to be detected according to the molybdenum plate defect distribution feature to obtain a molybdenum plate performance prediction sequence; Obtain a standard performance sequence of molybdenum plates, and obtain a performance requirement sequence of application scenarios of the molybdenum plates to be tested, compare the standard performance sequence of molybdenum plates with the predicted performance sequence of molybdenum plates, obtain a performance defect sequence of molybdenum plates, and obtain a quality inspection result of the molybdenum plates to be tested based on a judgment result of whether the performance defect sequence of the molybdenum plates meets the performance requirement sequence of application scenarios.

[0078] Specifically, the specific implementation method of the processor 10 for the above instructions can refer to Figures 1 to 3 The description of the relevant steps in the corresponding embodiments will not be repeated here.

[0079] Furthermore, if the module / unit integrated in the electronic device 1 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).

[0080] The present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor of an electronic device, the computer program can implement: Acquire a molybdenum plate to be inspected, and use a pre-built image device cluster to collect a visible light image and an ultrasonic image of the molybdenum plate to be inspected; According to a preset data processing strategy, the visible light image and the ultrasonic image are subjected to denoising to obtain denoised image data, a molybdenum plate defect recognition operation is performed on the denoised image data to obtain a surface defect area and an internal defect area, respectively, and a defect type marking operation is performed on the surface defect area and the internal defect area to obtain a surface defect type and an internal defect type, respectively, wherein the denoised image data includes a visible light denoised image and an ultrasonic denoised image; Using a pre-built augmented reality algorithm, the molybdenum plate defects are imaged according to the surface defect area, the surface defect type, the internal defect type and the internal defect area to obtain a molybdenum plate defect distribution image; Using a pre-trained molybdenum plate usability quality detection model to perform a feature extraction operation on the molybdenum plate defect distribution image to obtain a molybdenum plate defect distribution feature, and performing a performance prediction on the molybdenum plate to be detected according to the molybdenum plate defect distribution feature to obtain a molybdenum plate performance prediction sequence; Obtain a standard performance sequence of molybdenum plates, and obtain a performance requirement sequence of application scenarios of the molybdenum plates to be tested, compare the standard performance sequence of molybdenum plates with the predicted performance sequence of molybdenum plates, obtain a performance defect sequence of molybdenum plates, and obtain a quality inspection result of the molybdenum plates to be tested based on a judgment result of whether the performance defect sequence of the molybdenum plates meets the performance requirement sequence of application scenarios.

[0081] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, systems and methods can be implemented in other ways. For example, the system embodiments described above are only illustrative, and actual implementation may have other division methods.

[0082] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0083] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0084] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0085] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for detecting a molybdenum plate, characterized in that: The method comprises: Acquire a molybdenum plate to be inspected, and use a pre-built image device cluster to collect a visible light image and an ultrasonic image of the molybdenum plate to be inspected; According to a preset data processing strategy, the visible light image and the ultrasonic image are subjected to denoising to obtain denoised image data, a molybdenum plate defect recognition operation is performed on the denoised image data to obtain a surface defect area and an internal defect area, respectively, and a defect type marking operation is performed on the surface defect area and the internal defect area to obtain a surface defect type and an internal defect type, respectively, wherein the denoised image data includes a visible light denoised image and an ultrasonic denoised image; Using a pre-built augmented reality algorithm, the molybdenum plate defects are imaged according to the surface defect area, the surface defect type, the internal defect type and the internal defect area to obtain a molybdenum plate defect distribution image; Using a pre-trained molybdenum plate usability quality detection model to perform a feature extraction operation on the molybdenum plate defect distribution image to obtain a molybdenum plate defect distribution feature, and performing a performance prediction on the molybdenum plate to be detected according to the molybdenum plate defect distribution feature to obtain a molybdenum plate performance prediction sequence; Obtain a standard performance sequence of molybdenum plates, and obtain a performance requirement sequence of application scenarios of the molybdenum plates to be tested, compare the standard performance sequence of molybdenum plates with the predicted performance sequence of molybdenum plates, obtain a performance defect sequence of molybdenum plates, and obtain a quality inspection result of the molybdenum plates to be tested based on a judgment result of whether the performance defect sequence of the molybdenum plates meets the performance requirement sequence of application scenarios.

2. The method for detecting a molybdenum plate according to claim 1, wherein: According to the preset data processing strategy, the visible light image and the ultrasonic image are subjected to noise reduction processing to obtain noise reduction image data, the noise reduction image data is subjected to a molybdenum plate defect recognition operation to obtain a surface defect area and an internal defect area respectively, and the surface defect area and the internal defect area are subjected to a defect type marking operation to obtain a surface defect type and an internal defect type respectively, including: According to a preset data processing strategy, the visible light image is subjected to Gaussian filtering to obtain a visible light denoised image, and the ultrasonic image is subjected to filtering using a pre-constructed low-pass filter to obtain an ultrasonic denoised image; Defect recognition is performed on the visible light denoised image according to a pre-built Sobel operator to obtain surface defects of the molybdenum plate; Using a pre-built image segmentation algorithm to perform defect type identification and cutting operations on the surface defects of the molybdenum plate to obtain surface defect areas and surface defect types; Performing a waveform feature extraction operation based on a time domain signal on the ultrasonic image to obtain a time domain analysis result, and using a pre-built Fourier transform algorithm to transform the ultrasonic image into an ultrasonic frequency domain signal, and performing a frequency component analysis on the ultrasonic frequency domain signal to obtain a frequency domain analysis result; Defect identification is performed on the time domain analysis results and the frequency domain analysis results to obtain internal defect types and internal defect areas.

3. The detection method of molybdenum plate according to claim 2, characterized in that, The method of using the pre-built augmented reality algorithm to image the molybdenum plate defects according to the surface defect area, surface defect type, internal defect type and internal defect area to obtain a molybdenum plate defect distribution image includes: Using a pre-built augmented reality platform, constructing virtual objects for the surface defect area, surface defect type, internal defect type, and internal defect area to obtain a virtual defect object set, wherein the defect types of the virtual defect objects in the virtual defect object set are configured by different colors; Acquire a surrounding image of a pre-built augmented reality device, perform a molybdenum plate positioning operation based on the molybdenum plate to be detected on the surrounding image, and obtain a posture of the molybdenum plate; The molybdenum plate plane of the molybdenum plate posture is identified by using a pre-built plane detection algorithm, and the position and posture of the virtual defect object set is located according to the molybdenum plate plane to obtain a molybdenum plate defect distribution image.

4. The detection method of molybdenum plate according to claim 3, characterized in that, The method uses the pre-trained molybdenum plate usability quality detection model to perform feature extraction on the molybdenum plate defect distribution image to obtain the molybdenum plate defect distribution characteristics, and performs performance prediction on the molybdenum plate to be detected according to the molybdenum plate defect distribution characteristics to obtain the molybdenum plate performance prediction sequence, including: Using a pre-trained molybdenum plate usability quality detection model, a convolution operation is performed on the molybdenum plate defect distribution image to obtain a convolution matrix set; Performing a maximum pooling operation on the convolution matrix set to obtain a pooled feature set; Flattening the pooled feature set to obtain a defect distribution feature of the molybdenum plate, wherein the defect distribution feature of the molybdenum plate includes a defect area, a defect number, and a defect type; The defect distribution characteristics of the molybdenum plate are predicted based on a preset performance index set to obtain a molybdenum plate performance prediction sequence.

5. The detection method of molybdenum plate according to claim 4, characterized in that, Before performing the prediction operation on the defect distribution characteristics of the molybdenum plate based on a preset performance indicator set, the method further includes: Obtain a set of preset performance indicators including tensile strength, fatigue strength and electrical conductivity, wherein the tensile strength is expressed as: ， In the formula, represents the predicted tensile strength, Indicates standard tensile strength, It represents the influence coefficient of defect area on tensile strength. It represents the influence coefficient of the number of defects on the tensile strength, Indicates the influence coefficient of defect type on tensile strength, represents the defect area, represents the number of defects, Indicator of the impact of defect type; Wherein, the fatigue strength is expressed as: ， In the formula, represents the predicted fatigue strength, represents the standard fatigue strength, represents the mean stress, represents the influence coefficient of defect area on fatigue strength, represents the influence coefficient of the number of defects on fatigue strength, The influence coefficient of defect type on fatigue strength; Wherein, the electrical conductivity is expressed as: ， In the formula, represents the predicted conductivity, represents the standard conductivity, represents the influence coefficient of defect area on conductivity, represents the influence coefficient of the number of defects on the conductivity, represents the influence coefficient of defect type on conductivity, An exponential factor representing the effect of defects on conductivity.

6. The method for detecting a molybdenum plate according to claim 5, characterized in that: Before performing a feature extraction operation on the molybdenum plate defect distribution image using the pre-trained molybdenum plate availability quality detection model, the method further includes: Obtain a molybdenum plate usability quality detection model, and obtain a detection evaluation sample set, and allocate the detection evaluation sample set into a training set, a verification set, and a test set according to a preset allocation ratio, wherein the detection evaluation sample set includes: a sample molybdenum plate image and a true label sequence corresponding to the sample molybdenum plate image; According to the pre-configured training hyperparameters, a target sample is extracted from the training set in turn, and the molybdenum plate availability quality detection model is used to perform forward network prediction on the target sample to obtain a prediction performance index score sequence; Using a pre-built cross entropy loss algorithm, a loss value sequence between the true label sequence of the target sample and the prediction performance indicator score sequence is calculated; Minimize the loss value sequence to obtain a network model parameter sequence, and perform a reverse update of the parameters of the molybdenum plate availability quality detection model according to the network model parameter sequence to obtain an updated molybdenum plate availability quality detection model; Determine whether all the test evaluation samples in the training set participate in the training; When not all the detection evaluation samples in the training set participate in the training, return to the above step of extracting one target sample from the training set in sequence; When all the detection and evaluation samples in the training set participate in the training, the updated molybdenum plate availability quality detection model is evaluated using the verification set to obtain a verification and evaluation result, and according to the verification and evaluation result, the training hyperparameter is updated to obtain an updated hyperparameter; Using the test set to test the updated molybdenum plate availability quality detection model, to obtain a test accuracy rate; Determine whether the test accuracy is greater than a preset pass threshold; When the test accuracy is less than or equal to the qualified threshold, the hyperparameter is updated to the training hyperparameter, and the process returns to the step of allocating the detection evaluation sample set into a training set, a validation set, and a test set according to a preset allocation ratio; When the test accuracy is greater than the qualified threshold, the training process is stopped to obtain a trained molybdenum plate usability quality detection model.

7. The method for detecting a molybdenum plate according to claim 6, wherein: The molybdenum plate performance defect sequence is obtained by comparing the molybdenum plate standard performance sequence with the molybdenum plate performance prediction sequence, including: Performing difference calculation on the molybdenum plate standard performance sequence and the molybdenum plate performance prediction sequence to obtain a difference sequence set; Obtaining all performance labels whose difference scores are not zero in the difference sequence set to obtain a performance defect label set; The performance prediction score of each performance defect label in the performance defect label set in the molybdenum plate performance prediction sequence is extracted to obtain the molybdenum plate performance defect sequence.

8. The method for detecting a molybdenum plate according to claim 7, characterized in that: Before obtaining the molybdenum plate standard performance sequence, the method further includes: Comparing the pre-built minimum defect threshold sequence with the molybdenum plate performance prediction sequence to obtain a comparison result; According to the comparison result, it is determined whether there is a performance label in the molybdenum plate performance prediction sequence whose performance prediction score is less than the lowest defect threshold in the lowest defect threshold sequence; When there is a performance label in the molybdenum plate performance prediction sequence with a performance prediction score less than the lowest defect threshold in the lowest defect threshold sequence, the molybdenum plate to be tested is determined to be an unqualified product; When there is no performance tag in the molybdenum plate performance prediction sequence whose performance prediction score is less than the lowest defect threshold in the lowest defect threshold sequence, an operation of obtaining a standard performance sequence of the molybdenum plate is performed.

9. The method for detecting a molybdenum plate according to claim 8, wherein: After obtaining the quality test result of the molybdenum plate to be tested, the method further includes: Determining whether the quality inspection result is unqualified; When the quality inspection result is unqualified, the molybdenum plate type of the molybdenum plate to be inspected is obtained, wherein the molybdenum plate type includes a customized part molybdenum plate and a general size molybdenum plate; Determine whether the molybdenum plate type is a universal size molybdenum plate; When the molybdenum plate type is the universal size molybdenum plate, a maximum circumscribed matrix interception operation is performed on the molybdenum plate to be tested according to the application scenario performance requirement sequence to obtain a set of qualified molybdenum plates.

10. A molybdenum plate detection system, characterized in that: The system comprises: A molybdenum plate image acquisition module, used to acquire the molybdenum plate to be detected, and to use a pre-built image device cluster to acquire a visible light image and an ultrasonic image of the molybdenum plate to be detected; A defect image recognition module is used to perform noise reduction processing on the visible light image and the ultrasonic image according to a preset data processing strategy to obtain noise reduction image data, perform a molybdenum plate defect recognition operation on the noise reduction image data to obtain a surface defect area and an internal defect area respectively, and perform a defect type marking operation on the surface defect area and the internal defect area to obtain a surface defect type and an internal defect type respectively, wherein the noise reduction image data includes a visible light noise reduction image and an ultrasonic noise reduction image; A performance prediction module is used to use a pre-built augmented reality algorithm to perform molybdenum plate defect imaging according to the surface defect area, surface defect type, internal defect type and internal defect area to obtain a molybdenum plate defect distribution image, and use a pre-trained molybdenum plate usability quality detection model to perform a feature extraction operation on the molybdenum plate defect distribution image to obtain a molybdenum plate defect distribution feature, and perform performance prediction on the molybdenum plate to be detected according to the molybdenum plate defect distribution feature to obtain a molybdenum plate performance prediction sequence; The quality inspection and evaluation module is used to obtain the standard performance sequence of the molybdenum plate and the application scenario performance requirement sequence of the molybdenum plate to be inspected, compare the standard performance sequence of the molybdenum plate with the molybdenum plate performance prediction sequence, obtain the molybdenum plate performance defect sequence, and obtain the quality inspection result of the molybdenum plate to be inspected based on the judgment result of whether the molybdenum plate performance defect sequence meets the application scenario performance requirement sequence.

Images