A real-time detection system for defects in the target film of a neutron tube based on AI vision

Abstract

The present invention relates to the technical field of target film defect detection, and discloses a real-time detection system for neutron tube target film defects based on AI vision, including: a data acquisition module for obtaining the real image of the target film surface and manufacturing process parameters; a theoretical texture generation module for generating a theoretical texture image by using a trained generative adversarial network model; a spatio-temporal contrast analysis module for matching the theoretical texture image with the real image, extracting structural consistency features, texture direction consistency features and optical flow vector change features, and locating the mismatched area; a defect recognition module for constructing a texture evolution trend curve and comparing it with a preset multi-class defect evolution template curve to identify the defect type corresponding to the mismatched area; a process defect reverse inference module for predicting the type of process parameter deviation causing the defects in the mismatched area and generating a corresponding process adjustment strategy. The present invention realizes the real-time detection of neutron tube target film defects.

Description

Technical Field

[0001] The present invention belongs to the technical field of target film defect detection, and particularly relates to a real-time detection system for neutron tube target film defects based on AI vision. Background Art

[0002] The neutron tube is a key device in the radiation detection system, and the quality of its internal target film directly affects the neutron response sensitivity and device stability. The quality detection of existing neutron tube target films mostly relies on manual visual inspection or traditional image processing methods for defect identification.

[0003] The manual visual inspection method has problems such as low efficiency and strong subjectivity, and it is difficult to adapt to mass production; traditional image processing methods are mainly based on static features such as edge detection and gray distribution analysis, and it is difficult to accurately identify hidden defects with complex textures, weak defects, or caused by dynamic process disturbances. In addition, most of the existing detection methods only focus on the image itself, lack the association with the process parameters in the manufacturing process, and it is difficult to locate the cause of the defect and intelligently optimize the manufacturing parameters, and an effective feedback control mechanism cannot be formed. Summary of the Invention

[0004] The present invention provides a real-time detection system for neutron tube target film defects based on AI vision, which solves the technical problems of low accuracy, inability to identify defects caused by process disturbances, and lack of intelligent feedback ability linked to process parameters in the related art.

[0005] The present invention provides a real-time detection system for neutron tube target film defects based on AI vision, including:

[0006] A data acquisition module, configured to obtain real-time images of the target film surface and manufacturing process parameters during the manufacturing process of the neutron tube target film;

[0007] A theoretical texture generation module, configured to generate a theoretical texture image corresponding to the current manufacturing process parameters by using a trained generative adversarial network model based on the manufacturing process parameters;

[0008] A spatio-temporal contrast analysis module, configured to perform frame-level matching of the theoretical texture image and the real image in time according to the process time axis recorded during the manufacturing process, establish a spatial mapping relationship in combination with the target film manufacturing process trajectory model, and perform multi-scale difference calculation on the same region in the theoretical texture image and the real image, and extract structural consistency features, texture direction consistency features, and optical flow vector change features for locating the mismatch region;

[0009] A defect identification module, configured to receive the mismatch region, identify the defect type corresponding to the mismatch region by constructing a texture evolution trend curve and comparing the curve with a preset multi-class defect evolution template curve, and output the defect position, defect type, and confidence score;

[0010] A process defect back - inference module, which is used to reverse - predict the type of process parameter deviation that causes defects in the mismatch area based on the texture residual map of the mismatch area, in combination with the process - texture mapping model, and generate corresponding process adjustment strategies.

[0011] Further, the generator of the generative adversarial network model takes manufacturing process parameters as input. The manufacturing process parameters include: spraying speed, spin - coating speed, nozzle position, and temperature. The generator performs weighted fusion on the spatial distribution of the nozzle position and temperature through a spatio - temporal attention mechanism to generate a theoretical texture image that matches the spatial resolution of the target film manufacturing equipment.

[0012] The training data set of the generative adversarial network model includes simulated texture images generated by a physical simulation platform and real images collected during the actual manufacturing process.

[0013] Further, the discriminator of the generative adversarial network model is a multi - scale discriminator, and its input includes:

[0014] Multi - scale pyramids of the generated theoretical texture images and corresponding real images;

[0015] The difference vector between the current manufacturing process parameters and the preset process standard parameters, and the difference vector is converted into an attention mask through a fully - connected layer.

[0016] Further, the spatio - temporal contrast analysis module includes:

[0017] A time - axis synchronization unit, which is used to divide the theoretical texture image sequence and the real image sequence into multiple time periods according to the spraying trajectory based on the process time - axis and the process timestamps recorded by the manufacturing equipment. Each time period corresponds to one reciprocating motion cycle of the nozzle, and the frame indices of the theoretical texture image and the real image are aligned through timestamp interpolation.

[0018] A spatial mapping unit, which establishes an affine transformation matrix between the theoretical texture image coordinate system and the real image coordinate system based on the parameterized equation of the spraying path in the target film manufacturing process trajectory model, and performs alignment through bilinear interpolation.

[0019] A dynamic compensation unit, which adjusts the texture density distribution of the theoretical texture image according to the real - time change value of the spraying speed, so that the local texture density deviation between the spatially mapped theoretical texture image and the real image is less than the first preset threshold.

[0020] Further, the structural consistency feature is extracted according to the first registration region of the theoretical texture image and the real image in the spatial position. Among them, by calculating the ratio of the product of the brightness means of the theoretical texture image and the real image in the registration region to the sum of the squares of their brightness means, the first intermediate feature is obtained; by calculating the ratio of the product of the brightness standard deviations of the theoretical texture image and the real image in the registration region to the sum of the squares of their brightness standard deviations, the second intermediate feature is obtained; by calculating the ratio of the joint covariance of the theoretical texture image and the real image in the registration region to the product of their standard deviations, the third intermediate feature is obtained. Multiply the first intermediate feature, the second intermediate feature, and the third intermediate feature, and take the negative logarithm to obtain the structural consistency feature;

[0021] The texture direction consistency feature is extracted according to the local texture direction information of the theoretical texture image and the real image in the second registration region. Among them, the main texture directions of the theoretical texture image and the real image in the second registration region are respectively extracted by the structure tensor method, and the absolute value of the difference angle between the two is calculated to obtain the fourth intermediate feature. Map the fourth intermediate feature through the cosine function to obtain the texture direction consistency feature;

[0022] The optical flow vector change feature is extracted according to the local motion trend between consecutive frames of the theoretical texture image and the real image. Among them, the inter-frame optical flow vectors are respectively calculated for the theoretical texture image and the real image to obtain the theoretical texture optical flow vector and the real optical flow vector at the first target position. The ratio of the difference between the theoretical texture optical flow vector and the real optical flow vector to the modulus of the theoretical texture optical flow vector is used to obtain the optical flow vector change feature.

[0023] Further, the positioning of the mismatch region is based on the structural consistency feature, the texture direction consistency feature, and the optical flow vector change feature. The positioning process includes:

[0024] For each image block of the theoretical texture image and the real image, calculate three feature values respectively, and construct a three-channel texture difference feature map;

[0025] According to the preset feature deviation weight coefficient, perform weighted fusion on the three features to obtain the texture deviation score;

[0026] Mark the image blocks with the texture deviation score greater than the second preset threshold as mismatch regions.

[0027] Further, the defect recognition module constructs a texture evolution trend curve based on the image sequence extracted from the mismatch region. The texture evolution trend curve is based on the temporal features of the texture direction entropy and the contrast of the gray-level co-occurrence matrix, and uses the spraying speed as the weight coefficient for weighted modeling. The calculation formula of the texture evolution trend curve is: , where S represents the function value of the texture evolution trend curve, represents the spraying speed, represents a preset standard speed, represents the texture direction entropy of the current frame, t represents the index of the frame, represents the contrast of the gray-level co-occurrence matrix of the current frame, and respectively represent the maximum values of the texture direction entropy and the contrast of the gray-level co-occurrence matrix in the historical data.

[0028] Furthermore, the texture evolution trend curve of the mismatch region and the multi-class defect evolution template curves are respectively subjected to dynamic time warping distance calculation to obtain a matching distance. If the matching distance is less than the third preset threshold and the corresponding confidence score is greater than the fourth preset threshold, it is determined that the mismatch region is the corresponding defect type, where the confidence score is constructed based on the matching distance, and the defect types include: uneven sputtering, texture fracture, poor adhesion, and particle impurity deposition.

[0029] Furthermore, the process defect reverse inference module includes: a residual map segmentation unit, a parameter deviation calculation unit, and a deviation type judgment unit;

[0030] The residual map segmentation unit is used to divide the texture residual map of the mismatch region into multiple sub-regions according to the manufacturing process trajectory, and each sub-region corresponds to a control stage of the spraying path, where the texture residual map is obtained by performing pixel-by-pixel difference calculation on the mismatch region;

[0031] The parameter deviation calculation unit is used to reversely deduce the process parameter deviation amount that causes the abnormality based on the texture residual map of the mismatch region, including: extracting the texture residual values for each sub-region and calculating the deviation value of each manufacturing process parameter in combination with the sensitivity coefficient in the process-texture mapping model;

[0032] The deviation type determination unit is used to classify and judge the deviation amount of the manufacturing process parameters. When any deviation amount exceeds the fifth preset range, it is determined that the manufacturing process parameter is abnormal.

[0033] The beneficial effects of the present invention are as follows: By constructing a spatio-temporal registration model of the theoretical texture image and the real image and integrating multiple features such as structural consistency, direction consistency, and optical flow change, the present invention realizes high-precision identification of target film defects. At the same time, a process-texture mapping model is introduced to realize reverse inference of process parameter deviation based on the texture residual map, and then generate a process optimization strategy, which has the advantages of accurate defect identification, clear abnormal cause tracing, and intelligent manufacturing control, and significantly improves the manufacturing quality and stability of the target film. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a schematic diagram of the modules of a real-time detection system for target film defects of a neutron tube based on AI vision according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0035] Reference will now be made to example embodiments to discuss the subject matter described herein. It should be understood that discussing these embodiments is only to enable those skilled in the art to better understand and thus implement the subject matter described herein. Without departing from the scope of protection of the content of this specification, changes can be made to the functions and arrangements of the elements discussed. Each example can omit, substitute, or add various processes or components as needed. In addition, the features described for some examples can also be combined in other examples.

[0036] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in one or more embodiments of the present invention should have the ordinary meaning understood by those of ordinary skill in the art to which the present invention pertains. The "first", "second", and similar terms used in one or more embodiments of the present invention do not denote any order, quantity, or importance, but are only used to distinguish different components. Words such as "including" or "comprising" mean that the elements or objects appearing before this word cover the elements or objects listed after this word and their equivalents, without excluding other elements or objects. Words such as "connected" or "coupled" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Upper", "lower", "left", "right", etc. are only used to represent relative position relationships. When the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0037] As Figure 1 shown, a real-time detection system for neutron tube target film defects based on AI vision includes:

[0038] A data acquisition module 101 for real-time acquiring the real image of the target film surface and manufacturing process parameters during the manufacturing process of the neutron tube target film;

[0039] A theoretical texture generation module 102 for generating a theoretical texture image corresponding to the current manufacturing process parameters based on the manufacturing process parameters by using a trained generative adversarial network model;

[0040] A spatio-temporal contrast analysis module 103 for frame-level matching of the theoretical texture image and the real image in terms of time according to the process time axis recorded during the manufacturing process, establishing a spatial mapping relationship in combination with the target film manufacturing process trajectory model, and performing multi-scale difference calculation on the same region in the theoretical texture image and the real image to extract structural consistency features, texture direction consistency features, and optical flow vector change features for locating the mismatch region;

[0041] A defect recognition module 104, configured to receive the mismatch region, identify the defect type corresponding to the mismatch region by constructing a texture evolution trend curve and comparing the curve with a preset multi-class defect evolution template curve, and output the defect location, defect type, and confidence score;

[0042] A process defect reverse inference module 105, configured to reverse predict the type of process parameter deviation that causes the defect in the mismatch region based on the texture residual map of the mismatch region, and generate a corresponding process adjustment strategy.

[0043] In an embodiment of the present invention, the generative adversarial network model includes: a generator and a discriminator;

[0044] The generator of the generative adversarial network model takes manufacturing process parameters as input. The manufacturing process parameters include: spraying speed, spin coating speed, nozzle position, and temperature. The generator performs weighted fusion on the spatial distribution of the nozzle position and temperature through a spatio-temporal attention mechanism to generate a theoretical texture image that matches the spatial resolution of the target film manufacturing equipment;

[0045] Among them, the spraying speed is calculated by a linear encoder installed on the spraying slide rail in combination with the feedback signal of the click controller, and the unit of the spraying speed is mm / s; the spin coating speed is collected by the drive controller of the spin coating platform, and the unit of the spin coating speed is rps; the nozzle position is obtained by acquiring the coordinates of the nozzle in three-dimensional space; the temperature is collected by arranging a thermocouple array in the target film spraying area to obtain the temperature in the manufacturing area;

[0046] The training dataset of the generative adversarial network model includes simulated texture images generated by a physical simulation platform and real images collected during the actual manufacturing process.

[0047] In an embodiment of the present invention, during the model training stage, simulation images of the target film spraying process generated by a numerical simulation platform such as COMSOL are introduced to construct a texture evolution simulation diagram containing various typical process parameter combinations; and combined with the images of the actual production process to form a mixed dataset for training a conditional generative adversarial network, so that the generated theoretical texture image has higher reliability in terms of physical consistency and structural fidelity.

[0048] In an embodiment of the present invention, the discriminator of the generative adversarial network model is a multi-scale discriminator, and its input includes:

[0049] Multi-scale pyramids of the generated theoretical texture image and the corresponding real image. Specifically, the generated theoretical texture image and the real image are respectively downsampled by a Gaussian pyramid to generate image pyramids with different resolutions, which are respectively input into the sub-discriminators of the corresponding scales to achieve multi-level feature extraction from the global structure to the local details;

[0050] The difference vector between the current manufacturing process parameters and the preset process standard parameters, which is converted into an attention mask through a fully connected layer to enhance the discriminator's sensitivity to the area where the process parameters deviate. Specifically, calculate the difference vector between the current process parameters and the preset process standard parameters, input the difference vector into the fully connected layer, and then convert it into a spatial attention map through a Reshape operation to generate the attention mask.

[0051] In an embodiment of the present invention, the spatio-temporal contrast analysis module includes:

[0052] A time axis synchronization unit for dividing the theoretical texture image sequence and the real image sequence into multiple time periods according to the process time stamps of the manufacturing equipment along the spraying trajectory. Each time period corresponds to one reciprocating motion cycle of the spray head, and the frame indices of the theoretical texture image and the real image are aligned through timestamp interpolation;

[0053] A spatial mapping unit for accurately registering the theoretical texture image and the real image spatially, constructing a parametric equation of the spraying path based on the target film manufacturing process trajectory, and establishing an affine transformation matrix from the theoretical texture image coordinate system to the real image coordinate system. To avoid interpolation errors, bilinear interpolation is used to reconstruct the gray values of non-integer pixel points during the coordinate mapping process to obtain the mapped theoretical texture image;

[0054] Among them, the parametric equation of the spraying path is expressed as: , represents the position of the spray head, t is the process time stamp, and respectively represent the coordinate positions of the spray head in the x and y axis directions at time t;

[0055] By extracting the position, spraying direction, and scanning length of the spray head recorded by the manufacturing equipment, an affine transformation matrix from the theoretical image coordinate system to the real image coordinate system is constructed and represented in the following form: , where M is a 3×3 affine transformation matrix, a and b are the coordinate positions of the spray head in the x and y axis directions of the real image coordinate system, and u and v are the coordinate positions of the spray head in the x and y axis directions of the theoretical image coordinate system;

[0056] A dynamic compensation unit adjusts the texture density distribution of the theoretical texture image according to the real-time change value of the spraying speed, so that the local texture density deviation between the spatially mapped theoretical texture image and the real image is less than the first preset threshold, reducing the misregistration rate caused by process fluctuations in the image texture and improving the contrast accuracy.

[0057] In one embodiment of the present invention, the structural consistency feature is extracted according to the first registration region of the theoretical texture image and the real image in the spatial position. Wherein, by calculating the ratio of the product of the brightness means of the theoretical texture image and the real image in the registration region to the sum of the squares of their brightness means, a first intermediate feature is obtained; by calculating the ratio of the product of the brightness standard deviations of the theoretical texture image and the real image in the registration region to the sum of the squares of their brightness standard deviations, a second intermediate feature is obtained; by calculating the ratio of the joint covariance of the theoretical texture image and the real image in the registration region to the product of their standard deviations, a third intermediate feature is obtained. Multiply the first intermediate feature, the second intermediate feature and the third intermediate feature, and take the negative logarithm to obtain the structural consistency feature;

[0058] The calculation formula of the structural consistency feature is: , where represents the structural consistency feature, which is used to measure the overall consistency of the theoretical texture image and the real image in the local structural form, and assist in identifying the macroscopic texture mismatch caused by the fluctuation of the spraying thickness or the regional deformation; and respectively represent the brightness means of the theoretical texture image and the real image in the first registration region; and respectively represent the brightness standard deviations of the theoretical texture image and the real image in the first registration region; represents the covariance of the theoretical texture image and the real image in the first registration region; , and respectively represent the first constant, the second constant and the third constant, which are used to avoid the denominator being zero;

[0059] The texture direction consistency feature is extracted according to the local texture direction information of the theoretical texture image and the real image in the second registration region. Wherein, the main texture directions of the theoretical texture image and the real image in the second registration region are respectively extracted by the structure tensor method, and the absolute value of the included angle difference between them is calculated to obtain a fourth intermediate feature. The fourth intermediate feature is mapped through the cosine function to obtain the texture direction consistency feature;

[0060] The calculation formula of the texture direction consistency feature is: , where represents the texture direction consistency feature, which is used to detect the deviation degree of the theoretical texture and the real texture in the main direction, and locate the directional defects caused by the abnormal spin coating trajectory or the nozzle yaw; and respectively represent the main texture directions of the real image and the theoretical texture image in the second registration region, which are calculated by the structure tensor method; represents and the absolute value of the difference;

[0061] The optical flow vector change feature is extracted based on the local motion trend between consecutive frames of the theoretical texture image and the real image. Specifically, the inter-frame optical flow vectors of the theoretical texture image and the real image are calculated respectively to obtain the theoretical texture optical flow vector and the real optical flow vector at the first target position. The optical flow vector change feature is obtained by the ratio of the difference between the theoretical texture optical flow vector and the real optical flow vector to the magnitude of the theoretical texture optical flow vector.

[0062] The calculation formula of the optical flow vector change feature is: , where represents the optical flow vector change feature, which is used to compare the deviation degree of the dynamic behavior between consecutive frames of the theoretical texture image and the real image, and capture the motion inconsistency caused by process anomalies such as sudden deposition and adhesion failure. represents the optical flow vector between consecutive frames of the real image. represents the optical flow vector between consecutive frames of the theoretical texture image. The optical flow vector refers to the motion vector of the same pixel point in the image changing with time frames, indicating the motion direction and speed of the pixel, and is obtained by calculating through the traditional optical flow algorithm. represents a very small positive number. represents the Euclidean norm of.

[0063] In an embodiment of the present invention, the positioning of the mismatch region is based on the structural consistency feature, the texture direction consistency feature and the optical flow vector change feature. The positioning process includes:

[0064] For each image block of the theoretical texture image and the real image, three feature values are calculated respectively, and a three-channel texture difference feature map is constructed, and the three features are used as the multi-dimensional descriptor of the image block.

[0065] According to the preset feature deviation weight coefficient, the three features are weighted and fused to obtain the texture deviation score, which is used to represent the comprehensive deviation degree of the image block in terms of structure, direction and dynamic behavior.

[0066] The image blocks with the texture deviation score greater than the second preset threshold are marked as mismatch regions.

[0067] Through this three-feature fusion judgment mechanism, the system can accurately identify the texture offset of the theoretical texture image and the real image in terms of structure, direction and dynamic behavior, and improve the robustness and accuracy of the mismatch region detection.

[0068] In one embodiment of the present invention, the defect recognition module constructs a texture evolution trend curve based on an image sequence extracted from a mismatch region. The texture evolution trend curve is based on the temporal features of texture direction entropy and gray-level co-occurrence matrix contrast, and uses the spraying speed as a weight coefficient for weighted modeling. The calculation formula of the texture evolution trend curve is: , where S represents the function value of the texture evolution trend curve, represents the spraying speed, represents the preset standard speed, with the unit of mm / s, represents the texture direction entropy of the current frame, t represents the index of the frame, represents the gray-level co-occurrence matrix contrast of the current frame, and respectively represent the maximum values of texture direction entropy and gray-level co-occurrence matrix contrast in historical data.

[0069] Different from traditional texture features that are only calculated based on the static image, the texture evolution trend curve constructed in the present invention combines the change of spraying speed in the manufacturing process. By dynamically adjusting the relative weights of texture direction entropy and gray-level co-occurrence matrix contrast, the trend curve has better time-process consistency when capturing the defect evolution pattern, thereby improving the accuracy and discriminability of defect type recognition.

[0070] In one embodiment of the present invention, the dynamic time warping distance is calculated respectively between the texture evolution trend curve of the mismatch region and multi-class defect evolution template curves to obtain the matching distance. If the matching distance is less than the third preset threshold and the corresponding confidence score is greater than the fourth preset threshold, it is determined that the mismatch region is the corresponding defect type. Among them, the confidence score is constructed based on the matching distance, and the defect types include: uneven sputtering, texture fracture, poor adhesion, and particle impurity deposition.

[0071] Specifically, the dynamic time warping method is used to calculate the temporal shape similarity between the texture evolution trend curve and each defect evolution template curve. The calculation formula of the matching distance is: , DTW represents the matching distance, min represents the operation of taking the minimum value, represents the i-th defect evolution template curve, and i represents the index of the defect evolution template curve;

[0072] The calculation formula of the confidence score is: , represents the confidence score of the i-th defect evolution template, represents the mean value of the matching distances of all defect evolution templates.

[0073] In one embodiment of the present invention, the process defect backtracking module includes: a residual map segmentation unit, a parameter deviation calculation unit, and a deviation type judgment unit;

[0074] The residual map segmentation unit is used to divide the texture residual map of the mismatch area into multiple sub - regions according to the manufacturing process trajectory. Each sub - region corresponds to a control stage of the spraying path. Among them, the texture residual map is obtained by performing pixel - by - pixel difference calculation on the mismatch area;

[0075] The process parameter deviation calculation unit is used to inversely deduce the deviation amount of the process parameters causing the abnormality based on the texture residual map of the mismatch area, including: extracting the texture residual values for each sub - region, and combining with the sensitivity coefficients in the process - texture mapping model to calculate the deviation values of each manufacturing process parameter;

[0076] Among them, the texture residual value is obtained by the pixel gray - level difference of this sub - region of the texture residual map. The process - texture mapping model is used to represent the relationship between manufacturing process parameters and texture, and includes the sensitivity coefficients of manufacturing process parameters in each sub - region. The sensitivity coefficient reflects the average response amplitude caused by the change of manufacturing process parameters to the texture direction consistency feature in the sub - region. The deviation value of the manufacturing process parameter is obtained by weighted accumulation of the texture residual values and the corresponding sensitivity coefficients of multiple sub - regions;

[0077] The deviation type determination unit is used to classify and judge the deviation amount of the manufacturing process parameters. When any deviation amount exceeds the fifth preset range, it is determined that the manufacturing process parameter is abnormal;

[0078] Among them, the fifth preset range is preferably set to the range of 3 times the standard deviation. The process parameter deviation type includes the deviated manufacturing process parameter and the deviation direction. For example, the process parameter deviation type includes but is not limited to: "spraying speed is too high", "temperature is too low", etc. According to the process parameter deviation type, corresponding process adjustment strategies are generated. The process adjustment strategies include but are not limited to: reducing the spraying speed by 5% and extending the downward section time; raising the temperature of the spraying temperature zone to 580 °C and extending the constant - temperature section; reducing the spraying speed and at the same time increasing the uniform rotation section time.

[0079] In an embodiment of the present invention, by dividing the mismatch area into sub - regions according to the manufacturing process trajectory and combining with the sensitivity coefficients in the process - texture mapping model, the accurate inverse deduction of the process parameter deviation is realized. This method can effectively identify the potential process abnormal types causing image defects, improve the interpretability and regulation response ability of the manufacturing process, and has high intelligent diagnosis and closed - loop optimization value.

[0080] The above describes the embodiments of the present invention, but the present invention is not limited to the above - mentioned specific embodiments. The above - mentioned specific embodiments are only illustrative and not restrictive. Under the inspiration of this embodiment, those of ordinary skill in the art can also make many forms, all of which fall within the protection scope of this embodiment.

Claims

1. A real-time detection system for neutron tube target film defects based on AI vision, characterized in that, Including: A data acquisition module, configured to obtain a real image of the target film surface and manufacturing process parameters in real time during the manufacturing process of the neutron tube target film; A theoretical texture generation module, configured to generate a theoretical texture image corresponding to the current manufacturing process parameters by using a trained generative adversarial network model based on the manufacturing process parameters; A spatio-temporal contrast analysis module, configured to perform frame-level matching of the theoretical texture image and the real image in terms of time according to the process time axis recorded during the manufacturing process, establish a spatial mapping relationship in combination with the target film manufacturing process trajectory model, and calculate multi-scale differences for the same region in the theoretical texture image and the real image, and extract structural consistency features, texture direction consistency features, and optical flow vector change features for locating the mismatched region; Wherein, the structural consistency feature is extracted according to the first registration region in the spatial position of the theoretical texture image and the real image. By calculating the ratio of the product of the brightness means of the theoretical texture image and the real image in the registration region to the sum of the squares of their brightness means, a first intermediate feature is obtained; by calculating the ratio of the product of the brightness standard deviations of the theoretical texture image and the real image in the registration region to the sum of the squares of their brightness standard deviations, a second intermediate feature is obtained; by calculating the ratio of the joint covariance of the theoretical texture image and the real image in the registration region to the product of the brightness standard deviation, a third intermediate feature is obtained. Multiply the first intermediate feature, the second intermediate feature, and the third intermediate feature, and take the negative logarithm to obtain the structural consistency feature; A defect recognition module, configured to receive the mismatched region, identify the defect type corresponding to the mismatched region by constructing a texture evolution trend curve and comparing the curve with a preset multi-class defect evolution template curve, and output the defect position, defect type, and confidence score; Wherein, the defect recognition module constructs a texture evolution trend curve based on the image sequence extracted from the mismatched region. The texture evolution trend curve is based on the temporal features of texture direction entropy and gray-level co-occurrence matrix contrast, and uses the spraying speed as a weight coefficient for weighted modeling; A process defect reverse inference module, configured to reverse predict the type of process parameter deviation causing the defect in the mismatched region based on the texture residual map of the mismatched region in combination with the process-texture mapping model, and generate a corresponding process adjustment strategy.

2. The real-time detection system for neutron tube target film defects based on AI vision according to claim 1, characterized in that, The generator of the generative adversarial network model takes the manufacturing process parameters as input. The manufacturing process parameters include: spraying speed, spin coating speed, nozzle position, and temperature. The generator performs weighted fusion on the spatial distribution of the nozzle position and temperature through a spatio-temporal attention mechanism to generate a theoretical texture image matching the spatial resolution of the target film manufacturing equipment; The training data set of the generative adversarial network model includes simulated texture images generated by a physical simulation platform and real images collected during the actual manufacturing process.

3. The real-time detection system for neutron tube target film defects based on AI vision according to claim 1, wherein The discriminator of the generative adversarial network model is a multi-scale discriminator, and its input includes: A multi-scale pyramid of the generated theoretical texture image and the corresponding real image; A difference vector between the current manufacturing process parameters and the preset process standard parameters, and the difference vector is converted into an attention mask through a fully connected layer.

4. An in-situ detection system for neutron tube target membrane defects based on AI vision according to claim 1, characterized in that, The spatio-temporal contrast analysis module includes: A time-axis synchronization unit, which is used to divide a theoretical texture image sequence and a real image sequence into multiple time periods according to a spraying trajectory based on a process time axis and process timestamps recorded by manufacturing equipment. Each time period corresponds to a reciprocating motion cycle of a nozzle, and the frame indices of the theoretical texture image and the real image are aligned through timestamp interpolation; A spatial mapping unit, which establishes an affine transformation matrix between a theoretical texture image coordinate system and a real image coordinate system based on a parameterized equation of a spraying path in a target film manufacturing process trajectory model, and performs alignment through bilinear interpolation; A dynamic compensation unit, which adjusts the texture density distribution of a theoretical texture image according to a real-time change value of a spraying speed, so that the local texture density deviation between the spatially mapped theoretical texture image and the real image is less than a first preset threshold.

5. An in-situ detection system for defects of the neutron tube target film based on AI vision according to claim 1, characterized in that, The texture direction consistency feature is extracted according to local texture direction information of a theoretical texture image and a real image in a second registration region. Among them, the main texture directions of the theoretical texture image and the real image in the second registration region are respectively extracted through a structure tensor method, and the absolute value of the included angle difference between the two is calculated to obtain a fourth intermediate feature, and the fourth intermediate feature is mapped through a cosine function to obtain the texture direction consistency feature; The optical flow vector change feature is extracted according to the local motion trend between consecutive frames of a theoretical texture image and a real image. Among them, the inter-frame optical flow vectors of the theoretical texture image and the real image are respectively calculated to obtain a theoretical texture optical flow vector and a real optical flow vector at a first target position, and the ratio of the difference between the theoretical texture optical flow vector and the real optical flow vector to the modulus of the theoretical texture optical flow vector is used to obtain the optical flow vector change feature.

6. The real-time detection system for neutron tube target film defects based on AI vision according to claim 5, characterized in that, The positioning of a mismatched region is based on a structure consistency feature, a texture direction consistency feature, and an optical flow vector change feature. The positioning process includes: For each image block of a theoretical texture image and a real image, three eigenvalue are respectively calculated, and a three-channel texture difference feature map is constructed; According to a preset feature deviation weight coefficient, the three features are weighted and fused to obtain a texture deviation score; The image blocks with a texture deviation score greater than a second preset threshold are marked as mismatched regions.

7. The real-time detection system for neutron tube target film defects based on AI vision according to claim 2, characterized in that, The calculation formula of the texture evolution trend curve is as follows: , where S represents the function value of the texture evolution trend curve, represents the spraying speed, represents the preset standard speed, represents the texture direction entropy of the current frame, t represents the index of the frame, represents the contrast of the gray-level co-occurrence matrix of the current frame, and respectively represent the maximum values of the texture direction entropy and the contrast of the gray-level co-occurrence matrix in the historical data.

8. An in-situ detection system for neutron tube target membrane defects based on AI vision according to claim 1, characterized in that, The texture evolution trend curve of a mismatched region and the multi-class defect evolution template curves are respectively calculated for the dynamic time warping distance to obtain a matching distance. If the matching distance is less than a third preset threshold and the corresponding confidence score is greater than a fourth preset threshold, it is determined that the mismatched region is the corresponding defect type. Among them, the confidence score is constructed according to the matching distance, and the defect types include: uneven sputtering, texture fracture, poor adhesion, and particle impurity deposition.

9. The real-time detection system for neutron tube target film defects based on AI vision according to claim 1, characterized in that, The process defect backtracking module includes: a residual map segmentation unit, a parameter deviation calculation unit, and a deviation type judgment unit; The residual map segmentation unit is used to divide the texture residual map of a mismatched region into multiple sub-regions according to a manufacturing process trajectory. Each sub-region corresponds to a control stage of a spraying path. Among them, the texture residual map is obtained by performing a pixel-by-pixel difference calculation on the mismatched region; The parameter deviation calculation unit is used to inversely deduce the process parameter deviation amount that causes the abnormality based on the texture residual map of the mismatch region, including: extracting texture residual values for each sub-region, and calculating the deviation value of each manufacturing process parameter in combination with the sensitivity coefficient in the process-texture mapping model; The deviation type determination unit is used to classify and judge the deviation amount of the manufacturing process parameter. When any deviation amount exceeds the fifth preset range, it is determined that the manufacturing process parameter is abnormal.

Images