An image recognition-based service film material aging detection system

Abstract

The application discloses a kind of based on image recognition's service film material aging detection system, it is related to image recognition and computer vision technical field, including service image acquisition module, image normalization processing module, aging apparent feature construction module, historical change correlation module, aging candidate area identification module, aging grade evaluation module, membrane surface risk backwrite module.In the application, through the Chan-Vese candidate area identification mechanism under the constraint of historical change, the corresponding content after arrangement can participate in the initial level set profile formation, the apparent content with continuous change relation participates in aging area attraction, and the local apparent content lacking continuous change relation and forming abnormal deviation discriminant result is regarded as interference abnormal content and participates in area evolution constraint, so as to improve the aging candidate area boundary identification precision, enhance the level set profile convergence stability, and improve the exclusion ability to non-aging abnormal area.

Description

Technical Field

[0001] This invention relates to the fields of image recognition and computer vision technology, and in particular to an image recognition-based aging detection system for in-service membrane materials. Background Technology

[0002] During long-term use, membrane materials are susceptible to aging phenomena such as color changes, texture coarsening, crack propagation, and boundary damage due to the influence of light, temperature, humidity, external force, and pollutants. Existing detection methods usually collect images of membrane materials in service, divide the membrane surface area in the images, normalize the processing, and extract the appearance features. Then, based on the color, texture, crack, or boundary changes, the aging area and the degree of aging are determined, thereby helping maintenance personnel understand the service status of the membrane material.

[0003] Existing image recognition-based aging detection methods for in-service membrane materials rely heavily on differences in color, texture, or crack appearance in a single image, making it difficult to maintain the continuity and consistency of the membrane surface area's location and connectivity. Under interference from sudden changes in illumination, localized stains, or boundary damage, non-aging anomalies are easily misjudged as aging areas. Furthermore, the lack of correlation verification with historical membrane surface changes leads to unstable convergence of aging boundaries, making it difficult to accurately rewrite the aging level and risk range back to the corresponding membrane surface area. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing an image recognition-based aging detection system for service membrane materials.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an image recognition-based aging detection system for service membrane materials, comprising: a service image acquisition module, which receives service membrane material image data, performs membrane surface area processing on the service membrane material image data, and generates membrane surface area image data.

[0006] The image normalization processing module receives image data of the membrane surface area, performs service image normalization processing on the image data of the membrane surface area, and generates service membrane material normalized image data.

[0007] The aging appearance feature construction module receives normalized image data of the in-service membrane material, performs aging appearance feature construction processing on the normalized image data of the in-service membrane material, and generates aging appearance feature data.

[0008] The historical change association module receives aging appearance feature data and historical membrane surface image data, performs historical change association processing on the aging appearance feature data and historical membrane surface image data, and generates aging change association data.

[0009] The aging candidate region identification module receives aging appearance feature data and aging change correlation data, performs aging candidate region identification processing based on the Chan-Vese segmentation algorithm on the aging appearance feature data and aging change correlation data, and generates aging candidate region data.

[0010] The aging level assessment module receives aging candidate area data and aging change correlation data, performs aging level assessment processing on the aging candidate area data and aging change correlation data, and generates aging level assessment data.

[0011] The membrane risk write-back module receives aging level assessment data and membrane area image data, performs membrane risk write-back processing on the aging level assessment data and membrane area image data, and generates aging test results for the in-service membrane material.

[0012] As a further description of the above technical solution: the steps of the service image acquisition module to generate membrane surface area image data include: the service image acquisition module receives service membrane material image data, arranges the image content in the service membrane material image data according to the acquisition order, and generates the sequentially arranged image content.

[0013] The image content after being arranged in sequence is divided into the membrane area content and the background area content, and the membrane area content is retained.

[0014] Connect the edge-continuous content and the internal continuous content in the membrane area according to the membrane area to generate continuous membrane area content.

[0015] The content of the continuous membrane surface area is divided into regions to generate multiple membrane surface regions, and each membrane surface region is used as a region location.

[0016] Preserve the positional relationship of each membrane region within the content of a continuous membrane region, and connect membrane regions that are adjacent in position to generate region connection relationships.

[0017] The membrane surface area content, location, and connectivity of each membrane surface area are connected to generate membrane surface area image data.

[0018] As a further description of the above technical solution: the steps of the image normalization processing module to generate normalized image data of the service membrane material include: the image normalization processing module receives the membrane surface area image data, and arranges the membrane surface area content, area position and area connection relationship of each membrane surface area according to the area position, so that the content of each membrane surface area forms a continuous arrangement relationship in the area position.

[0019] The image brightness and darkness representations in each membrane area are collected according to their location, and the image brightness and darkness representations of adjacent membrane areas are compared according to their region connectivity. The brightness and darkness differences obtained from the comparison are applied to the corresponding membrane area content, so that the content of adjacent membrane areas forms a continuous brightness and darkness transition along the region connectivity.

[0020] The image color representations in the membrane area content after the completion of continuous light and dark transitions are aggregated according to the region location, and the image color representations of adjacent membrane area content are compared according to the region connection relationship. The color differences obtained from the comparison are applied to the corresponding membrane area content, so that the content of adjacent membrane area content forms a continuous color transition along the region connection relationship.

[0021] The local light and dark abrupt changes in the content of the membrane area after the continuous color transition are connected with the continuous light and dark transitions of the adjacent membrane area content, so that the local light and dark abrupt changes are incorporated into the light and dark continuity of the corresponding area position.

[0022] After completing the connection of light and dark continuity, the content of the membrane area is aligned according to the region position, so that the content of the membrane area corresponding to the same membrane area falls into the same region position, and the correspondence between the aligned membrane area content and the region connection relationship is maintained.

[0023] Connect the contents of adjacent membrane areas with continuous regional connectivity to the boundary, so that the contents of adjacent membrane areas are included in the same membrane area.

[0024] The content, location, and connection relationships of the membrane area after completing continuous light and dark transitions, continuous color transitions, continuous light and dark relationship connections, area position alignment, and boundary connection are encapsulated to generate normalized image data of the membrane material in service.

[0025] As a further description of the above technical solution: the steps of the aging appearance feature construction module to generate aging appearance feature data include: the aging appearance feature construction module receives normalized image data of the service membrane material, and arranges the membrane surface area content, area position and area connection relationship in the normalized image data of the service membrane material according to the area position, so that the content of each membrane surface area forms a continuous connection relationship within the same membrane surface area.

[0026] The image color expressions in each membrane area are aggregated according to their location, and the image color expressions within the same membrane area and between adjacent membrane areas are compared according to the region connection relationship. The membrane area content where the color expression shows continuous differences is marked as color change content.

[0027] The image brightness and color representations in each membrane area are combined according to their location. The corresponding content after combination is then compared for continuity based on the region connection relationship. The corresponding content that appears repeatedly in adjacent regions and is arranged in a continuous direction is marked as texture change content.

[0028] The corresponding content that extends continuously in the texture change content is connected to the color change content in terms of position, and the extension relationship between adjacent film areas is verified along the region connection relationship. The corresponding content that crosses the position of adjacent areas and maintains continuous extension is marked as crack continuous content.

[0029] The continuous edge content in the membrane area is compared with the region connection relationship, and the corresponding content that is broken, misaligned or missing in the continuous edge content is marked as boundary damaged content.

[0030] The content of color change, texture change, crack continuity, and boundary damage is connected according to the region position, so that each corresponding content is consistent with the position of the same film area.

[0031] The color change content, texture change content, crack continuity content, and boundary damage content that have been connected at the completed location are encapsulated according to the content of the film surface area to generate aging appearance feature data.

[0032] As a further description of the above technical solution: the steps of the historical change association module to generate aging change association data include: the historical change association module receives aging appearance feature data and historical film surface image data, and arranges the color change content, texture change content, crack continuity content and boundary damage content in the aging appearance feature data according to the regional position, so that each corresponding content is consistent with the position of the film surface area.

[0033] Historical membrane surface image data is mapped to aging appearance feature data according to regional location, so that the image content corresponding to the same membrane surface area in the historical membrane surface image data falls into the same regional location, and the image content in the same regional location has corresponding image color expression, image brightness expression, edge continuity content and regional connection relationship.

[0034] By comparing the color representation and color change content of the same membrane area in historical membrane image data, the continuity relationship of color change of the same membrane area between historical membrane image data and aging appearance feature data is determined.

[0035] By comparing the brightness and color representations of the same membrane area in historical membrane image data with the texture changes, the continuity relationship of texture changes in the same membrane area between historical membrane image data and aging appearance feature data is determined.

[0036] By connecting the texture variation continuity relationship with the crack continuity content according to the region connection relationship, the crack variation continuity relationship that spans adjacent regions and maintains extension between historical film surface image data and aging appearance feature data is determined.

[0037] By comparing the continuous edge content and boundary damage content corresponding to the same membrane area in historical membrane image data according to the regional connectivity relationship, the boundary damage continuity relationship of the same membrane area between historical membrane image data and aging appearance feature data is determined.

[0038] The relationships of color change continuity, texture change continuity, crack change continuity, and boundary damage continuity are connected according to their regional locations to form the change association content corresponding to the same film area.

[0039] The change-related content is positionally linked with the color change content, texture change content, crack continuity content, and boundary damage content in the aging appearance feature data to generate aging change-related data.

[0040] As a further description of the above technical solution: the steps of the aging candidate region identification module to generate aging candidate region data include: the aging candidate region identification module receives aging appearance feature data and aging change association data, and connects and arranges the color change content, texture change content, crack continuity content, boundary damage content, and change association content corresponding to the same film surface area according to the region position.

[0041] The corresponding content after arrangement is imported into the level set initialization process of the Chan-Vese segmentation algorithm to form an initial level set outline within the region location, and to maintain the positional correspondence between the initial level set outline and the aging appearance feature data and aging change correlation data.

[0042] Divide the area into inner and outer regions along the initial horizontal set outline. Collect the color change content, texture change content, crack continuity content, and boundary damage content corresponding to the inner region, and collect the color change content, texture change content, crack continuity content, and boundary damage content corresponding to the outer region.

[0043] Match the corresponding content in the inner and outer regions with the change-related content of the same membrane region. Write the content of regions with continuous change relationship into the inner evolution direction of the level set, and write the content of regions without continuous change relationship into the outer exclusion direction of the level set.

[0044] The color change content, texture change content, crack continuity content, and boundary damage content in the evolution direction inside the horizontal set are uniformly merged, so that the appearance content with continuous change relationship participates in the attraction of aging region.

[0045] The corresponding content in the exclusion direction outside the horizontal set is input into the local Reed-Xiaoli anomaly detection algorithm to form a local neighborhood around the region. The local distribution relationship between color change content, texture change content and boundary damage content in the local neighborhood is calculated. The deviation degree between the content corresponding to the position to be judged in the exclusion direction outside the horizontal set and the local distribution relationship is calculated to form an anomaly deviation judgment result. Local appearance content that lacks continuous change relationship and forms an anomaly deviation judgment result is regarded as interference anomaly content.

[0046] The aging region attracts and interferes with anomalous content, which is incorporated into the region evolution process of the Chan-Vese segmentation algorithm. This causes the level set contour to advance towards regions with continuous aging changes and exit from regions corresponding to the interference anomalous content.

[0047] The contours of the level set are continuously evolved according to the Chan-Vese segmentation algorithm, and the contour movement content and change-related content in adjacent evolution sequences are checked for correspondence.

[0048] Retain the outline content that continuously corresponds to the content associated with the change, and remove the outline content whose correspondence is interrupted, so that the outline of the level set converges to the aging boundary position.

[0049] The continuous connected film surface regions within the converged aging boundary are merged, and the merged region content is connected with the corresponding color change content, texture change content, crack continuity content, boundary damage content, and change association content to generate aging candidate region data.

[0050] As a further description of the above technical solution: the steps of the local Reed-Xiaoli anomaly detection algorithm are as follows: arrange the corresponding content in the exclusion direction outside the horizontal set according to the region position, so that the color change content, texture change content, boundary damage content and change-related content maintain the correspondence in the same region position.

[0051] A local neighborhood is defined around the position to be judged in the direction of exclusion outside the horizontal set. The color change content, texture change content, and boundary damage content adjacent to the position to be judged within the local neighborhood are then collected.

[0052] The color change content, texture change content, and boundary damage content gathered in the local neighborhood are grouped according to the same region location, and the local distribution relationship between the color change content, texture change content, and boundary damage content in the local neighborhood is calculated.

[0053] The Reed-Xiaoli deviation of the content corresponding to the position to be judged in the exclusion direction outside the horizontal set is calculated with respect to the local distribution relationship, forming the abnormal deviation judgment result of the content corresponding to the position to be judged relative to the local neighborhood.

[0054] The continuity relationship between the abnormal deviation judgment result and the change-related content corresponding to the position to be judged is checked to determine whether the content corresponding to the position to be judged is missing a continuous change relationship.

[0055] Local appearance content that lacks continuous change relationships and forms abnormal deviations from the discrimination results is regarded as interfering abnormal content.

[0056] As a further description of the above technical solution: the steps of the aging level assessment module to generate aging level assessment data include: the aging level assessment module receives aging candidate region data and aging change correlation data, connects the aging candidate regions in the aging candidate region data with the change correlation content corresponding to the same membrane surface region in the aging change correlation data according to the region location, and generates a level assessment object.

[0057] The continuity of color changes in the assessment objects is judged, and the extension range of color changes in the same aging candidate area, the connection between adjacent areas, and the corresponding changes are compared to form the basis for color change level.

[0058] The coarsening degree of texture changes in the grade assessment object is judged, and the recurrence range, arrangement direction continuity and corresponding change-related content of texture changes in the same aging candidate area are compared to form the basis for texture change grade.

[0059] The extent of expansion of the crack continuity content in the grade assessment object is judged. The extension relationship of the crack continuity content across adjacent areas, the positional connection relationship between the crack continuity content and the color change content and texture change content, and the corresponding change-related content are compared to form the basis for crack continuity grade.

[0060] The degree of damage to the boundary damage content in the assessment object is judged. The extent of the breakage, misalignment or missing content at the boundary of the aging candidate area and the corresponding changes are compared to form the basis for the boundary damage level.

[0061] The color change level criteria, texture change level criteria, crack continuity level criteria, and boundary damage level criteria are connected according to the region location, and the level criteria that are continuously associated with the change are retained in the same aging candidate region to generate the region level criteria.

[0062] Based on the combination of color changes, texture changes, crack continuity, and boundary damage in the regional grading criteria, each aging candidate region is graded, and aging grade content is generated.

[0063] The aging level information is linked with the corresponding aging candidate areas, area locations, and change-related information to generate aging level assessment data.

[0064] As a further description of the above technical solution: the steps of the membrane risk write-back module to generate the aging test results of the service membrane material include: the membrane risk write-back module receives aging level assessment data and membrane area image data, arranges the aging level content, aging candidate area, area location and change association content in the aging level assessment data according to the area location, and matches them with the membrane area content and area connection relationship in the membrane area image data to generate write-back positioning content.

[0065] The aging level content in the write-back positioning content is written into the film surface area content corresponding to the same area location, so that each aging candidate area has the corresponding aging level content in the film surface area image data, thus generating the level write-back content.

[0066] The aging candidate area in the graded write-back content is connected to the membrane area with the region connection relationship. The membrane area content that crosses the adjacent area position and has continuous aging grade content is connected into the same write-back range to generate the risk write-back range.

[0067] The aging level content and change-related content in the risk write-back scope are verified to correspond. The aging level content that continuously corresponds to the change-related content is retained in the corresponding write-back scope, and the aging level content with interrupted correspondence is separated from the corresponding write-back scope to generate verified risk write-back content.

[0068] The verified risk write-back content is linked with the image content in the membrane area content, so that the aging candidate area, aging level content, area location and corresponding image content maintain a corresponding relationship under the same membrane area, and image-related write-back content is generated.

[0069] The image-associated write-back content is embedded into the membrane area image data according to the regional location and regional connection relationship, so that the risk write-back range and the corresponding membrane area content are kept in the same position, thus generating membrane risk write-back content.

[0070] The aging candidate area, aging level, area location, area connection relationship, change-related content and corresponding image content in the membrane risk write-back content are encapsulated to generate the aging test results of the membrane material in service.

[0071] The present invention has the following beneficial effects: 1. In the present invention, through the membrane surface region position continuity mechanism, the image content in the image data of the service membrane material can form a continuous data foundation through membrane surface region content, continuous membrane surface region content, region position and region connection relationship, thereby improving the membrane surface region positioning capability, reducing the influence of background region content on subsequent detection, and making subsequent normalization processing, aging appearance feature construction, candidate region identification and risk write-back have a consistent position foundation; through the cross-region normalization and appearance feature consistency mechanism, the image brightness and darkness expression, image color expression, local brightness and darkness abrupt change content, boundary connection, as well as color change content, texture change content, crack continuity content and boundary damage content can all maintain correspondence along the region position and region connection relationship, thereby improving the continuity and interpretability of aging appearance feature data, and enhancing the recognition sensitivity of color change, texture change, crack continuity and boundary damage.

[0072] 2. In this invention, a historical change correlation verification mechanism enables aging appearance feature data to form color change continuity, texture change continuity, crack change continuity, and boundary damage continuity relationships with historical film surface image data in the same film surface area and location. This improves the accuracy of identifying real aging changes, reduces misjudgments caused by changes in illumination, local stains, or instantaneous appearance anomalies, and enhances the traceability of aging detection results. Furthermore, a Chan-Vese candidate region identification mechanism under historical change constraints allows the corresponding arranged content to participate in the initial level set contour formation, enabling appearance content with continuous change relationships to participate in the aging region absorption. This approach introduces localized appearances lacking continuous variation and exhibiting abnormal deviations from the discrimination results as interfering anomalies, thus contributing to regional evolution constraints. This improves the accuracy of aging candidate region boundary recognition, enhances the convergence stability of level set contours, and improves the ability to exclude non-aging anomaly regions. Through a grade assessment and membrane risk write-back mechanism, the aging grade content can maintain correspondence with aging candidate regions, region locations, variation-related content, membrane surface region content, and region connectivity. This allows the aging test results of in-service membrane materials to simultaneously express the aging grade, risk write-back range, and corresponding image content, improving the localization capability, interpretability, and decision support capability of the test results. Attached Figure Description

[0073] Figure 1 This is a structural diagram of the in-service membrane aging detection system based on image recognition of the present invention. Detailed Implementation

[0074] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0075] Reference Figure 1 The present invention provides an embodiment of an image recognition-based aging detection system for service membrane materials, comprising: a service image acquisition module, which receives service membrane material image data, performs membrane surface area processing on the service membrane material image data, and generates membrane surface area image data.

[0076] The image normalization processing module receives image data of the membrane surface area, performs service image normalization processing on the image data of the membrane surface area, and generates service membrane material normalized image data.

[0077] The aging appearance feature construction module receives normalized image data of the in-service membrane material, performs aging appearance feature construction processing on the normalized image data of the in-service membrane material, and generates aging appearance feature data.

[0078] The historical change association module receives aging appearance feature data and historical membrane surface image data, performs historical change association processing on the aging appearance feature data and historical membrane surface image data, and generates aging change association data.

[0079] The aging candidate region identification module receives aging appearance feature data and aging change correlation data, performs aging candidate region identification processing based on the Chan-Vese segmentation algorithm on the aging appearance feature data and aging change correlation data, and generates aging candidate region data.

[0080] The aging level assessment module receives aging candidate area data and aging change correlation data, performs aging level assessment processing on the aging candidate area data and aging change correlation data, and generates aging level assessment data.

[0081] The membrane risk write-back module receives aging level assessment data and membrane area image data, performs membrane risk write-back processing on the aging level assessment data and membrane area image data, and generates aging test results for the in-service membrane material.

[0082] In this embodiment, the service image acquisition module receives service membrane material image data, maps the image content in the service membrane material image data to the acquisition order, arranges the image content acquired earlier in the acquisition order before the image content acquired later in the acquisition order, sets the image content acquired in the acquisition order as image content with adjacent order, and arranges all image content in a continuous arrangement along the acquisition order, so that the image content in the service membrane material image data is transformed from an unarranged state to image content with an acquisition order succession relationship, generating image content in the order arrangement.

[0083] The sequentially arranged image content is divided into membrane surface area content and background area content. Content in the sequentially arranged image content that corresponds to the membrane surface range and can participate in the formation of the membrane surface area is classified as membrane surface area content, while content in the sequentially arranged image content that does not correspond to the membrane surface range and does not participate in the formation of the membrane surface area is classified as background area content. The membrane surface area content is retained from the sequentially arranged image content, and the retained membrane surface area content continues to maintain the arrangement relationship formed by the acquisition order.

[0084] The continuous edge content and the continuous internal content in the membrane surface area are connected according to the membrane surface range. The continuous edge content is connected along the extension relationship of the membrane surface range, so that the continuous edge content distributed at the edge of the membrane surface range forms a continuous state. The continuous internal content is connected along the distribution relationship within the membrane surface range, so that the continuous internal content forms a continuous state. Then, the continuous edge content and the continuous internal content are connected within the same membrane surface range, so that the continuous edge content and the continuous internal content together constitute a continuous membrane surface area content, generating continuous membrane surface area content.

[0085] The content of the continuous membrane surface area is divided into regions. Content that is continuous in position and corresponds to the same membrane surface range is assigned to the same membrane surface area, while content that is separated in position and corresponds to different membrane surface ranges is assigned to different membrane surface areas. This divides the continuous membrane surface area content into multiple distinguishable membrane surface areas. Each membrane surface area is a region location, so that each region location corresponds to a membrane surface area, and multiple membrane surface areas jointly contain the continuous membrane surface area content.

[0086] The positional relationship of each membrane region within the continuous membrane region content is preserved, and the relative position of each membrane region within the continuous membrane region content is retained as the positional relationship between each membrane region; membrane regions with adjacent positions are connected, and the membrane regions corresponding to adjacent positions are connected according to the positional relationship, so that the membrane regions with adjacent positions can form a connection that can carry the content of the continuous membrane region, thus generating a region connection relationship.

[0087] The content, location, and connectivity of each membrane region are connected. The content of the membrane region within the same region is connected to the corresponding location, allowing the content to determine its position across multiple membrane regions. The connectivity of the location is then linked to the same membrane region, creating a correspondence between the content, location, and connectivity within the same region. By connecting the content, location, and connectivity of each membrane region, membrane region image data is generated.

[0088] In this embodiment, the image normalization processing module receives the membrane area image data, connects the membrane area content of each membrane area in the membrane area image data with the corresponding area position, so that the membrane area content of each membrane area falls into the corresponding area position; retains the area connection relationship corresponding to each membrane area under the corresponding area position, so that the area connection relationship can point to the membrane area content of adjacent positions; then arranges the membrane area content along the area position, so that the membrane area content with adjacent area positions remains adjacent in the arrangement state, and the membrane area content corresponding to the area connection relationship remains continuous in the arrangement state, thereby forming a continuous arrangement relationship of the membrane area content in the area position.

[0089] The image brightness and darkness representations in each membrane surface region are aggregated according to their location, concentrating the image brightness and darkness representations within the same membrane surface region to the corresponding location, thus enabling the image brightness and darkness representations in different locations to form a correspondence along the region connection relationship. The image brightness and darkness representations in membrane surface regions with adjacent region connection relationships are compared to determine the brightness and darkness differences between adjacent membrane surface regions. The brightness and darkness differences obtained from the comparison are applied to the corresponding membrane surface region content, causing the image brightness and darkness representations of membrane surface regions with more prominent brightness and darkness to connect with those of adjacent membrane surface regions, and causing the image brightness and darkness representations of membrane surface regions with weaker brightness and darkness to transition to those of adjacent membrane surface regions, thereby creating a continuous brightness and darkness transition between adjacent membrane surface regions along the region connection relationship.

[0090] After a continuous transition between light and dark, the image color expressions in the membrane area are aggregated according to their location, concentrating the image color expressions within the same membrane area to their corresponding location. This ensures that adjacent membrane areas can correspond in their image color expressions. The image color expressions in adjacent membrane areas are compared to determine the color differences between them. These color differences are then applied to the corresponding membrane areas, causing the prominent color expressions to connect with the image color expressions of adjacent membrane areas, while adjusting the color expressions of membrane areas that deviate from the norm to a continuous state along the regional connection relationship. This results in a continuous color transition between adjacent membrane areas along the regional connection relationship.

[0091] The local light-dark transition content in the membrane area after continuous color transition is connected with the continuous light-dark transition of the adjacent membrane area content, so that the local light-dark transition content remains in the corresponding area position; the local light-dark transition content is connected with the continuous light-dark transition already formed around the corresponding area position, so that the local light-dark transition content does not break away from the light-dark continuity relationship between the adjacent membrane area content; then the local light-dark transition content is incorporated into the light-dark continuity relationship of the corresponding area position, so that the membrane area content after continuous color transition still maintains the light-dark continuity with the adjacent membrane area content at the local light-dark transition position.

[0092] After completing the connection of light and dark continuity, the content of the membrane area is aligned according to the area position, so that the content of the membrane area corresponding to the same membrane area is grouped to the same area position; the content of the membrane area that deviates from the corresponding area position is adjusted to the corresponding area position, so that the content of the membrane area corresponding to the same membrane area is consistent in the area position; during the alignment process, the correspondence between the content of the membrane area and the area connection relationship is preserved, so that the area connection relationship can still connect the content of the membrane area with the adjacent positions, and the aligned content of the membrane area continues to follow the area connection relationship.

[0093] The content of adjacent membrane areas with continuous regional connection relationships is connected by boundary, so that the content of adjacent membrane areas close to the boundary of the regional connection relationship is connected to each other; the content of adjacent membrane areas pointed to by the regional connection relationship is continuously connected along the boundary, so that the boundary between the content of adjacent membrane areas is not broken; then the content of adjacent membrane areas with completed boundary connection is connected into the same membrane area, so that the content of adjacent membrane areas within the same membrane area remains continuous at the boundary, and the regional connection relationship continues to correspond to the content of adjacent membrane areas with completed boundary connection.

[0094] After completing continuous light and dark transitions, continuous color transitions, continuous light and dark relationship connections, region position alignment, and boundary connections, the content, region position, and region connection relationships of the membrane surface region are encapsulated. The content of the membrane surface region under the same membrane surface region is connected to the corresponding region position, so that the content of the membrane surface region can be determined by the region position. The region connection relationships under the corresponding region position are connected to the same membrane surface region, so that the content, region position, and region connection relationships of the membrane surface region remain corresponding under the same membrane surface region. Then, the content of each membrane surface region is uniformly encapsulated along the region position and region connection relationship, so that the membrane surface region content that has completed the normalization process can remain continuous within the same membrane surface range, generating normalized image data of the service membrane material.

[0095] In this embodiment, the aging appearance feature construction module receives normalized image data of the service membrane material, connects the membrane surface area content in the normalized image data of the service membrane material with the corresponding area position, so that the content of each membrane surface area falls under the corresponding area position; the area connection relationship is retained between adjacent area positions, so that the content of adjacent membrane surface areas can form a continuity along the area connection relationship; the content of each membrane surface area is arranged according to the area position, so that the content of membrane surface areas with adjacent area positions remains adjacent in the arrangement order, and the content of membrane surface areas corresponding to the area connection relationship remains connected after arrangement, thereby forming a continuous continuity relationship of each membrane surface area within the same membrane surface range.

[0096] The image color expressions in each membrane surface region are aggregated according to their region location, so that the image color expressions within the same membrane surface region are concentrated under the corresponding region location, and the image color expressions between adjacent membrane surface regions can form a correspondence along the region connection relationship; the image color expressions within the same membrane surface region are compared sequentially to determine the continuous differences between consecutive image color expressions within the same membrane surface region; the image color expressions between adjacent membrane surface regions are compared sequentially along the region connection relationship to determine the continuous differences in color expressions across adjacent regions; the membrane surface region content where the color expression shows continuous differences is linked to the corresponding region location and marked as color change content.

[0097] The image brightness and color representations in each membrane area are combined according to their location, so that the image brightness and color representations in the same membrane area form corresponding content at the corresponding location. The combined corresponding content is then compared continuously along the region connection relationship, so that the corresponding content that appears repeatedly in the same membrane area is connected, the corresponding content that appears repeatedly in adjacent regions is connected, and the corresponding content that is continuously arranged in the region connection relationship is maintained. The corresponding content that appears repeatedly in adjacent regions and is continuously arranged in the region connection relationship is maintained and marked as texture change content.

[0098] The corresponding content that extends continuously within the texture change content is positionally connected to the color change content, ensuring that the texture change content and color change content remain corresponding in the same area. The extension relationship of the texture change content between adjacent film areas is verified along the area connection relationship, ensuring that the texture change content that crosses adjacent areas can maintain continuous extension with the texture change content in the adjacent film area. The texture change content that maintains continuous extension is then positionally connected to the color change content in the same area, thus identifying the corresponding content that crosses adjacent areas and maintains continuous extension, and marking it as crack continuity content.

[0099] The continuous edge content within the membrane surface area is compared with the region connection relationship to ensure that the continuous edge content corresponds to the region connection relationship between adjacent membrane surface areas at the corresponding region position. Content that fails to continue along the region connection relationship within the continuous edge content is identified as disconnected content. Content that deviates from the region connection relationship of adjacent membrane surface areas within the continuous edge content is identified as misaligned content. Content that fails to connect with the region connection relationship of adjacent membrane surface areas within the continuous edge content is identified as missing content. Content that is disconnected, misaligned, or missing is kept connected to the corresponding region position and marked as boundary damaged content.

[0100] Color variation content, texture variation content, crack continuity content, and boundary damage content are connected according to their regional positions, ensuring that the color variation content, texture variation content, crack continuity content, and boundary damage content are all in the same position with their corresponding film surface areas. Furthermore, color variation content, texture variation content, crack continuity content, and boundary damage content within the same film surface area are interconnected, creating a consistent positional relationship among the corresponding contents within the same film surface area, and enabling each corresponding content to collectively support the film surface area content within the same film surface area.

[0101] The color change content, texture change content, crack continuity content, and boundary damage content that have been completed and connected are encapsulated according to the content of the film surface area, so that the color change content, texture change content, crack continuity content, and boundary damage content in the same film surface area content remain corresponding; the encapsulated color change content, texture change content, crack continuity content, and boundary damage content in each film surface area content are connected according to the region position, so that the corresponding content in different film surface areas can be maintained along the region connection relationship; the encapsulated color change content, texture change content, crack continuity content, and boundary damage content that have been completed and connected along the region position are output as the same data result to generate aging appearance feature data.

[0102] In this embodiment, the historical change association module receives aging appearance feature data and historical film surface image data. It arranges the color change content in the aging appearance feature data according to its regional location, so that the color change content corresponds to its corresponding film surface region; it arranges the texture change content according to its regional location, so that the texture change content corresponds to its corresponding film surface region; it arranges the crack continuity content according to its regional location, so that the crack continuity content corresponds to its corresponding film surface region; it arranges the boundary damage content according to its regional location, so that the boundary damage content corresponds to its corresponding film surface region; then it connects the color change content, texture change content, crack continuity content, and boundary damage content under the same film surface region, so that each corresponding content maintains a consistent position within the same film surface region, and forms an arrangement state in which the corresponding content in the aging appearance feature data can be regionally corresponding to the historical film surface image data.

[0103] Historical membrane surface image data and aging appearance feature data are mapped according to their regional locations. Image content in the historical membrane surface image data corresponding to the same membrane surface area is connected to the same regional location in the aging appearance feature data, so that the image content in the historical membrane surface image data corresponds to the color change content, texture change content, crack continuity content, and boundary damage content in the aging appearance feature data in terms of regional location. At the same regional location, the image content in the historical membrane surface image data is connected to the image color expression, the image content in the historical membrane surface image data is connected to the image brightness and darkness expression, the image content in the historical membrane surface image data is connected to the edge continuity content, and the image content in the historical membrane surface image data is connected to the regional connection relationship, so that the image content at the same regional location has corresponding image color expression, image brightness and darkness expression, edge continuity content, and regional connection relationship.

[0104] The image color representation and color change content corresponding to the same membrane area in historical membrane image data are compared and connected. The image color representation within the same membrane area is connected to the color change content at the corresponding location, so that the image color representation and color change content correspond to each other at the same location. The continuity of image color representation and color change content is compared along the location of the same membrane area, so that the image color representation in the historical membrane image data and the color change content in the aging appearance feature data form a correspondence within the same membrane area. The image color representation and color change content between adjacent locations are compared along the region connection relationship, so that the image color representation and color change content across adjacent locations form a continuous correspondence, thus determining the continuity relationship of color change of the same membrane area between historical membrane image data and aging appearance feature data.

[0105] In historical membrane image data, the image brightness and color representations corresponding to the same membrane area are compared and connected. Image brightness and color representations are linked to texture changes within the same region, ensuring positional correspondence among these elements within the same membrane area. Furthermore, the continuity of these elements is compared along regional positions within the same membrane area, establishing a sequential correspondence between historical membrane image data and aging appearance feature data regarding texture changes. Finally, the image brightness, color, and texture changes between adjacent regions are compared along regional connectivity, ensuring continuous correspondence across adjacent regions. This determines the continuity of texture changes for the same membrane area between historical membrane image data and aging appearance feature data.

[0106] The texture variation continuity relationship and the crack continuity content are connected according to the region connection relationship. The content of the texture variation continuity relationship located in the same film area is connected to the crack continuity content in the corresponding region position, so that the texture variation continuity relationship and the crack continuity content remain corresponding in the same region position. The texture variation continuity relationship that crosses the adjacent region position is connected to the crack continuity content along the region connection relationship, so that the texture variation continuity relationship between adjacent region positions can carry the crack continuity content. Then, the content that maintains the extension between the historical film image data and the aging appearance feature data is connected to the crack continuity content, so that the extension state that crosses the adjacent region position remains continuous under the region connection relationship. The crack variation continuity relationship that crosses the adjacent region position and maintains the extension between the historical film image data and the aging appearance feature data is determined.

[0107] In historical membrane image data, the continuous edge content and boundary damage content corresponding to the same membrane area are compared according to the region connection relationship. The continuous edge content in the historical membrane image data is connected to the boundary damage content in the same region, so that the continuous edge content and boundary damage content form a positional correspondence in the same membrane area. The continuity state of the continuous edge content and boundary damage content is compared along the regional position in the same membrane area, so that the continuous edge content in the historical membrane image data and the boundary damage content in the aging appearance feature data form a front-to-back correspondence. The continuous edge content and boundary damage content between adjacent regions are compared along the region connection relationship, so that the continuous edge content and boundary damage content across adjacent regions maintain a continuous correspondence, thus determining the boundary damage continuity relationship of the same membrane area between historical membrane image data and aging appearance feature data.

[0108] Color change continuity, texture change continuity, crack change continuity, and boundary damage continuity are connected according to their regional positions. Color change continuity in the same regional position is connected to the same film surface region, texture change continuity in the same regional position is connected to the same film surface region, crack change continuity in the same regional position is connected to the same film surface region, and boundary damage continuity in the same regional position is connected to the same film surface region. Then, color change continuity, texture change continuity, crack change continuity, and boundary damage continuity that can be inherited along the regional connection relationship in adjacent regional positions are connected, so that each continuity relationship remains corresponding within the same film surface region and between adjacent regional positions, forming the change association content corresponding to the same film surface region.

[0109] The change-related content is positionally linked with the color change content, texture change content, crack continuity content, and boundary damage content in the aging appearance feature data. The change-related content is connected to the color change content in the same area, thus giving the color change content a corresponding change-related content; the change-related content is connected to the texture change content in the same area, thus giving the texture change content a corresponding change-related content; the change-related content is connected to the crack continuity content in the same area, thus giving the crack continuity content a corresponding change-related content; the change-related content is connected to the boundary damage content in the same area, thus giving the boundary damage content a corresponding change-related content; finally, the positionally linked change-related content and the corresponding content in the aging appearance feature data are kept together in the same film surface area to generate aging change-related data.

[0110] In this embodiment, the aging candidate region identification module receives aging appearance feature data and aging change correlation data. It connects the color change content, texture change content, crack continuity content, and boundary damage content in the aging appearance feature data to the region positions, ensuring that the color change content, texture change content, crack continuity content, and boundary damage content remain corresponding within the same film surface region. It also connects the change correlation content corresponding to the same film surface region in the aging change correlation data to the same region position, ensuring that the change correlation content is consistent with the color change content, texture change content, crack continuity content, and boundary damage content in terms of region position. Finally, it arranges the connected corresponding content along the region positions, forming an arrangement state within the same film surface region that can be input into the subsequent level set initialization process, thus forming the arranged corresponding content. The formula for forming the arranged corresponding content is: ; ; ; Indicates the location of the area The following is the appearance content, which consists of color change content, texture change content, crack continuity content, and boundary damage content. Indicates the location of the area The following content consists of color change content, texture change content, crack continuity content, boundary damage content, and related change content. This represents the corresponding content after arranging the elements along the regional positions. , , , , Representing the location of the area respectively The following includes content related to color changes, texture changes, crack continuity, boundary damage, and related changes. This represents the set of all regional locations.

[0111] The corresponding content after arrangement is imported into the level set initialization process of the Chan-Vese segmentation algorithm. Color change content, texture change content, crack continuity content, boundary damage content, and change-related content are used together as the positional basis for forming the initial level set contour. Within the region, the initial level set contour is formed around the arranged corresponding content, maintaining a positional correspondence between the initial level set contour and the color change content, texture change content, crack continuity content, and boundary damage content in the aging appearance feature data. Simultaneously, the initial level set contour maintains a positional correspondence with the change-related content in the aging change-related data, forming an initial level set contour with a positional correspondence between it and the aging appearance feature data and the aging change-related data. The formula for forming the initial level set contour is: , ; Represents the initial level set outline. This represents the outline boundary formed around the corresponding content after arrangement. Indicates the initial level set contour in the region location. The outline state below, Indicates the location of the area The corresponding content below, This indicates the corresponding content after sorting. This represents the set of all regional locations.

[0112] The initial horizontal set outline is divided into inner and outer regions along its outline position. Color variations, texture variations, crack continuity, and boundary damage content located in the inner region of the initial horizontal set outline are grouped into the inner region, while the same content located in the outer region is grouped into the outer region. During the grouping process, the content corresponding to each item in the inner region retains its original region position, and the content corresponding to each item in the outer region retains its original region position, thus completing the grouping of corresponding content in the inner and outer regions. The formula for dividing the inner and outer regions is as follows: ; ; ; ; This represents the inner region obtained by dividing the initial level set contour. This represents the outer region obtained by dividing the initial level set contour. This indicates the color changes, texture changes, crack continuity, and boundary damage within the inner region. This indicates the color changes, texture changes, crack continuity, and boundary damage in the outer region. Indicates the location of the area The following appearance content, Indicates the initial level set contour in the region location. The outline state below.

[0113] The corresponding content in the inner and outer regions is matched with the change-related content of the same film surface region. The color change content, texture change content, crack continuity content, and boundary damage content in the inner region are connected with the change-related content at the same region location. Similarly, the color change content, texture change content, crack continuity content, and boundary damage content in the outer region are connected with the change-related content at the same region location. The evolution direction of the inner side of the level set is written into the formula: The direction of exclusion outside the level set is written into the formula: ; This represents the region of content written into the evolution direction inside the level set. This indicates that the content to be written should be in the region outside the horizontal set, excluding certain directions. Indicates the location of the area The apparent content and the content related to the changes below have a continuous changing relationship. Indicates the location of the area The apparent content and the related content lack a continuous relationship of change. Indicates the location of the area The changes are related to the content below. and These represent the inner and outer regions, respectively.

[0114] Regions exhibiting continuous change relationships are identified and written into the inner evolution direction of the level set. Regions lacking continuous change relationships are identified and written into the outer exclusion direction of the level set, ensuring that the inner and outer exclusion directions of the level set respectively receive the matching results between the inner and outer regions and the change-related content. The aging region attraction formation formula is as follows: ; Indicates the location of the area Attracting aging areas below, This indicates the consistency merging of color variation content, texture variation content, crack continuity content, and boundary damage content within the same area. This indicates that the apparent content and the related content have a continuous changing relationship. This indicates that the apparent content and the related content lack a continuous relationship of change. This indicates that the area is not involved in the attraction of aging areas.

[0115] The color change content, texture change content, crack continuity content, and boundary damage content within the inner evolution direction of the horizontal set are uniformly merged. Color change content and texture change content in the same region are connected; crack continuity content is connected to color change content and texture change content in the same region; and boundary damage content is connected to color change content, texture change content, and crack continuity content in the same film area. Then, appearance content with continuous change relationships is maintained along its region position within the inner evolution direction of the horizontal set, allowing this appearance content to participate in aging region attraction. Aging region attraction represents the inward pushing effect of appearance content with continuous change relationships in the inner evolution direction of the horizontal set on the horizontal set contour, causing the horizontal set contour to move towards the region where color change content, texture change content, crack continuity content, and boundary damage content continuously correspond, and retaining this region as part of the aging candidate region. The formula for arranging content corresponding to the outer exclusion direction of the horizontal set is as follows: ; ; Indicates the location of the region in the direction of exclusion outside the horizontal set. The corresponding content below, This represents the corresponding content after excluding the direction of the horizontal set and arranging it according to the region position. , , , Representing the location of the area respectively The following content includes color changes, texture changes, boundary damage, and related changes. This indicates the direction of exclusion outside the level set.

[0116] The local Reed-Xiaoli anomaly detection algorithm receives corresponding content from the exclusion direction outside the horizontal set. It connects color change content from this direction to the corresponding region, ensuring the color change content retains its original position within the region. It also connects texture change content to the corresponding region, ensuring a correspondence between texture change content and color change content within the same region. Furthermore, it connects boundary damage content to the corresponding region, ensuring a correspondence between boundary damage content and both color and texture change content within the same region. Finally, it connects related change content to the corresponding region, ensuring a correspondence between related change content and color, texture, and boundary damage content within the same region. The algorithm then arranges these connected elements along the region, maintaining the correspondence between them and ensuring the corresponding content from the exclusion direction outside the horizontal set can continue to be processed around the location to be determined.

[0117] Based on the fact that the corresponding content in the exclusion direction outside the horizontal set has maintained its correspondence according to its regional position, a local neighborhood is defined around the position to be determined in the exclusion direction outside the horizontal set. Color variation content located in the region adjacent to the position to be determined is gathered into the local neighborhood, ensuring that the local neighborhood can contain color variation content adjacent to the position to be determined. Texture variation content located in the region adjacent to the position to be determined is gathered into the same local neighborhood, ensuring that the local neighborhood can contain texture variation content adjacent to the position to be determined. Boundary breakage content located in the region adjacent to the position to be determined is gathered into the same local neighborhood, ensuring that the local neighborhood can contain boundary breakage content adjacent to the position to be determined. Then, the color variation content, texture variation content, and boundary breakage content adjacent to the position to be determined within the local neighborhood are kept in their corresponding regional positions, so that each content within the local neighborhood maintains both its adjacency to the position to be determined and its correspondence within its regional position. Local neighborhood formation formula: ; , Indicates the location to be determined. The defined local neighborhood, This represents the color variation content, texture variation content, and boundary damage content gathered within a local neighborhood. Indicates the position to be determined. This indicates the location of the region adjacent to the location to be determined. This indicates the direction of exclusion outside the level set.

[0118] The color change content, texture change content, and boundary damage content, after being aggregated within a local neighborhood, are combined according to their location within the same region. Color change content and texture change content within the same region are connected, creating corresponding combinations within the local neighborhood. Boundary damage content within the same region is also connected to color change content, ensuring correspondence between them. Furthermore, the combined color change content, texture change content, and boundary damage content at different locations within the local neighborhood are interconnected, allowing these combinations to participate in the calculation of local distribution relationships. By calculating the corresponding states between color change content, texture change content, and boundary damage content within the local neighborhood, a local distribution relationship is formed among them.

[0119] Formula for calculating local distribution relationship: ; ; ; Indicates the location of the area The following is a local appearance content composed of color change content, texture change content, and boundary damage content. This represents the local distribution center of local appearance content within a local neighborhood. This indicates the local distribution relationship between color changes, texture changes, and boundary damage within a local neighborhood. This indicates the number of regions in a local neighborhood. Indicates the location to be determined. The defined local neighborhood, This indicates transpose and is used to form local distribution relationships.

[0120] The Reed-Xiaoli deviation calculation is performed on the content corresponding to the position to be judged in the exclusion direction outside the level set and its local distribution relationship. This involves mapping the color changes in the content corresponding to the position to be judged to the color changes in the local distribution relationship, ensuring a deviation relationship is formed between the color changes in the position to be judged and the color changes in the local neighborhood; mapping the texture changes in the content corresponding to the position to be judged to the texture changes in the local distribution relationship, ensuring a deviation relationship is formed between the texture changes in the position to be judged and the texture changes in the local neighborhood; mapping the boundary damage in the content corresponding to the position to be judged to the boundary damage in the local distribution relationship, ensuring a deviation relationship is formed between the boundary damage in the position to be judged and the boundary damage in the local neighborhood; and finally, the deviation between the content corresponding to the position to be judged and its local distribution relationship is calculated to determine any abnormal deviations of the content corresponding to the position to be judged relative to its local neighborhood, thus forming an abnormal deviation judgment result. The Reed-Xiaoli deviation calculation formula is as follows: ; ; Indicates the position to be determined The corresponding content's deviation from its local neighborhood in Reed-Xiaoli terms. This indicates the result of the abnormal deviation judgment. Indicates the position to be determined The following is a local appearance content composed of color change content, texture change content, and boundary damage content. This represents the local distribution center of local appearance content within a local neighborhood. The inverse matrix representing the local distribution relationship. This indicates transpose.

[0121] The continuity relationship between the anomaly deviation judgment result and the change-related content corresponding to the position to be judged is verified. The anomaly deviation judgment result is connected to the change-related content corresponding to the position to be judged, so that the anomaly deviation judgment result and the change-related content corresponding to the position to be judged can maintain a correspondence in the same area. The continuity relationship between the local appearance content corresponding to the anomaly deviation judgment result and the change-related content corresponding to the position to be judged is verified, so that the continuity relationship between the local appearance content and the change-related content is determined. When the content corresponding to the position to be judged cannot form a continuity relationship with the change-related content, it is determined that the content corresponding to the position to be judged lacks a continuity relationship. When the content corresponding to the position to be judged can form a continuity relationship with the change-related content, the content corresponding to the position to be judged is not considered to lack a continuity relationship. This determines whether the content corresponding to the position to be judged lacks a continuity relationship. Continuity relationship verification formula: ; This indicates the verification result regarding whether the content corresponding to the position to be judged lacks a continuous change relationship. This indicates that the content corresponding to the location to be determined lacks a continuous change relationship with the related content at that location. This indicates that the content corresponding to the location to be determined and the related content corresponding to the location to be determined have a continuous change relationship. Indicates the position to be determined The corresponding changes are related to the content. This indicates a lack of continuous change relationships. It indicates a continuous change relationship.

[0122] Local appearance content lacking continuous change relationships and forming anomaly deviation judgment results is considered as interfering anomaly content. This local appearance content lacking continuous change relationships is mapped to the anomaly deviation judgment results to ensure consistency. Local appearance content that simultaneously satisfies both lack of continuous change relationships and forming anomaly deviation judgment results is determined from the content corresponding to the position to be judged. This determined local appearance content is then used as interfering anomaly content, inheriting both the anomaly deviation judgment results and continuity verification results from the local Reed-Xiaoli anomaly detection algorithm. The formula for determining interfering anomaly content is: ; Indicates the position to be determined The corresponding interference anomaly content, Indicates the position to be determined The following are the local appearance contents. Indicates the position to be determined The abnormal deviation judgment results below This indicates the result of the continuity check. This indicates that the location to be determined does not contain any interfering or abnormal content.

[0123] The attraction of aging regions and the inclusion of interfering anomalous content are incorporated into the region evolution process of the Chan-Vese segmentation algorithm. The attraction of aging regions is connected to the inner evolution direction of the level set, causing the level set contour to advance along regions with continuous aging relationships. Interfering anomalous content is connected to the outer exclusion direction of the level set, causing the level set contour to exit from the region corresponding to the interfering anomalous content. During the region evolution process, the level set contour is simultaneously constrained by both the attraction of aging regions and the interference of anomalous content, completing both the advancement towards regions with continuous aging relationships and the exit from regions corresponding to the interference of anomalous content. The Chan-Vese region evolution process formula is as follows: ; ; Indicates the location of the area The repulsive effect corresponding to the abnormal interference content. Indicates the location of the area The following is interference and abnormal content. Indicates the first Regional location in each evolutionary sequence The corresponding level set contour state, Indicates the first Regional location in each evolutionary sequence The corresponding level set contour state, Indicates the location of the area Attracting aging areas below, Indicates the location of the area The interference and abnormal content below corresponds to the exclusion effect.

[0124] The level set contours evolve continuously according to the Chan-Vese segmentation algorithm. The position of the level set contour in each evolutionary order is mapped to the position of the level set contour in the adjacent evolutionary order, thus determining the contour movement content in the adjacent evolutionary order. The contour movement content is then matched with the change-related content at the same region location to ensure that the contour movement content corresponds to the change-related content, and this mapping verification between contour movement content and change-related content in adjacent evolutionary orders is completed. The mapping verification formula between contour movement content and change-related content is as follows: ; ; This represents the contour movement content in adjacent evolutionary sequences. This indicates the verification result of the correspondence between the moved content and the related changed content. Indicates the first The level set contour states under each evolutionary order Indicates the first The level set contour states under each evolutionary order Indicates the location of the area The changes are related to the content below. This indicates that the content of the outline movement corresponds continuously with the related content of the change. This indicates that the correspondence between the content that moved the outline and the content associated with the change was interrupted.

[0125] Contour content that continuously corresponds to the changed content is retained, allowing it to continue participating in level set contour convergence. Contour content with interrupted correspondence is removed, preventing it from participating in level set contour convergence. By retaining continuously corresponding contour content and removing those with interrupted correspondence, the level set contour gradually converges to the aging boundary position. Aging boundary position convergence formula: ; ; This represents the aging boundary after convergence. This indicates outline content that has been removed due to a break in the correspondence. Indicates the first Level set contours under each evolutionary order Indicates the location of the area The outline content and the related changes in the content correspond continuously. Indicates the location of the area The correspondence between the outline content and the related content of the changes is interrupted.

[0126] The content of continuously connected film surface regions within the converged aging boundary is merged. Film surface regions within the same area within the aging boundary are connected to the same aging boundary. Continuously connected film surface regions along the region position within the aging boundary are further merged. Then, the merged region content is connected with corresponding color change content, texture change content, crack continuity content, boundary damage content, and change-related content. This ensures that the merged region content has corresponding appearance content and change-related content at the same region position, generating aging candidate region data. The formula for generating aging candidate region data is: ; ; This represents the content of the continuously connected and merged region within the converged aging boundary. This represents aging candidate region data. This represents the aging boundary after convergence. , , , , Representing the location of the area respectively The following includes content related to color changes, texture changes, crack continuity, boundary damage, and related changes. Indicates the location of the area The content of the continuously connected region located within the aging boundary after convergence.

[0127] In this embodiment, the aging level assessment module receives aging candidate region data and aging change correlation data. It arranges the aging candidate regions in the aging candidate region data according to their regional locations, so that each aging candidate region can correspond to its corresponding regional location. It connects the change correlation content corresponding to the same membrane surface region in the aging change correlation data to the same regional location, so that the change correlation content can maintain a correspondence with the aging candidate region in terms of regional location. Then, it connects the aging candidate region data and the change correlation content in the same regional location, so that the aging candidate region data and the aging change correlation data together form a level assessment object that can be used for subsequent level assessment, and generates a level assessment object.

[0128] The continuity of color changes in the assessment objects is determined. The extension range of the color change content within the same aging candidate area is connected to the corresponding aging candidate area, so that the color change content can maintain the corresponding regional position within the aging candidate area. The connection relationship between the color change content in adjacent regional positions is compared with the corresponding change-related content, so that the extension range of the color change content and the connection relationship between adjacent regional positions jointly correspond to the change-related content. The color change content that has been compared is then used as the basis for the color change level assessment, forming the basis for the color change level assessment.

[0129] The texture changes in the assessment objects are coarsened to a certain degree. The recurrence range of the texture changes within the same aging candidate area is connected to the corresponding aging candidate area, so that the texture changes can maintain a corresponding position within the aging candidate area. The continuity of the arrangement direction of the texture changes is connected to the connection relationship between adjacent areas, so that the texture changes can form a continuous correspondence along the position. Then, the recurrence range, the continuity of the arrangement direction, and the corresponding change-related content are compared to ensure that the texture changes and change-related content maintain a correspondence within the same aging candidate area, thus forming the basis for the texture change level.

[0130] The degree of expansion of the crack continuity content in the grade assessment object is judged. The extension relationship of the crack continuity content across adjacent areas is connected with the corresponding aging candidate area, so that the crack continuity content can maintain continuity within the aging candidate area and between adjacent areas. The positional connection relationship between the crack continuity content and the color change content is compared, so that the crack continuity content and the color change content can maintain correspondence in the same area. The positional connection relationship between the crack continuity content and the texture change content is compared, so that the crack continuity content and the texture change content can maintain correspondence in the same area. Finally, the extension relationship of the crack continuity content across adjacent areas, the positional connection relationship with the color change content and the texture change content, and the corresponding change-related content are connected to form the basis for the crack continuity grade.

[0131] The degree of damage to boundary-damaged content within the assessment object is determined. The disconnected portion of the boundary-damaged content at the boundary of the aging candidate region is connected to the corresponding aging candidate region, ensuring the disconnected portion corresponds to the aging candidate region boundary. Similarly, the misaligned portion of the boundary-damaged content at the boundary of the aging candidate region is connected to the corresponding aging candidate region, ensuring the misaligned portion corresponds to the aging candidate region boundary. The missing portion of the boundary-damaged content at the boundary of the aging candidate region is also connected to the corresponding aging candidate region, ensuring the missing portion corresponds to the aging candidate region boundary. Finally, the disconnected, misaligned, or missing portions are compared with the corresponding change-related content to ensure the damage manifestation of the boundary-damaged content at the aging candidate region boundary corresponds to the change-related content, thus forming the basis for the boundary damage level.

[0132] Color change level criteria, texture change level criteria, crack continuity level criteria, and boundary damage level criteria are connected according to their regional locations. Color change level criteria, texture change level criteria, crack continuity level criteria, and boundary damage level criteria in the same regional location are connected to their corresponding aging candidate regions. Then, level criteria that continuously correspond to the change-related content are retained in the same aging candidate region, so that each level criterion can jointly bear the change-related content in the same aging candidate region to generate regional level criteria.

[0133] The region grading criteria are based on color variation, texture variation, crack continuity, and boundary damage, which are determined according to their combination relationships. The corresponding content of color variation in the region grading criteria is connected to the corresponding content of texture variation, allowing both color and texture variations to participate in the grading of each aging candidate region. Similarly, the corresponding content of crack continuity in the region grading criteria is connected to color variation and texture variation, ensuring that crack continuity corresponds to the same aging candidate region along with both color and texture variations. Furthermore, the corresponding content of boundary damage in the region grading criteria is connected to color variation, texture variation, and crack continuity, enabling boundary damage to form a combination relationship with other content that constitutes the aging candidate region. Finally, based on the combination relationships of color variation, texture variation, crack continuity, and boundary damage in the region grading criteria, each aging candidate region is graded, generating aging grade content.

[0134] The aging level content is connected to the corresponding aging candidate region, region location, and change-related content. The aging level content corresponding to each aging candidate region is connected to that aging candidate region, so that the aging level content can be mapped to a specific aging candidate region. The aging level content is connected to the corresponding region location, so that the aging level content can maintain positional correspondence with the aging candidate region through the region location. The aging level content is connected to the corresponding change-related content, so that the aging level content can inherit the change-related content in the aging change-related data. Finally, the connected aging level content, corresponding aging candidate region, region location, and change-related content are encapsulated together to generate aging level assessment data.

[0135] In this embodiment, the membrane risk write-back module receives aging level assessment data and membrane area image data. It connects the aging level content in the aging level assessment data to the corresponding area location, so that the aging level content can correspond to the same area location; it connects the aging candidate area to the corresponding area location, so that the aging candidate area can maintain correspondence with the aging level content; it connects the change-related content to the corresponding area location, so that the change-related content can maintain correspondence with the aging level content and the aging candidate area; then it matches the completed arrangement of aging level content, aging candidate area, area location, and change-related content with the membrane area content and area connection relationship in the membrane area image data, so that the membrane area content can receive the aging level content, aging candidate area, and change-related content at the same area location, and generates write-back positioning content.

[0136] The aging level content in the write-back positioning content is written into the membrane area content corresponding to the same area location. The aging level content is then connected to the membrane area content under the same area location in the write-back positioning content, so that the membrane area content can carry the aging level content. The aging candidate region is kept in correspondence with the membrane area content that has already been written with the aging level content, so that each aging candidate region has the corresponding aging level content in the membrane area image data. Finally, the write-back positioning content that has completed the writing of the aging level content is used as the level write-back content to generate the level write-back content.

[0137] In the aging candidate area of ​​the graded write-back content, the membrane surface area adjacent to the region connection relationship is judged for acceptance. The aging candidate area is connected to the membrane surface area content under the adjacent region position, so that the aging candidate area can correspond to the adjacent membrane surface area along the region connection relationship. The membrane surface area content that crosses the adjacent region position and is continuous in aging grade is connected, so that the membrane surface area content under the adjacent region position can form the same write-back range under the same aging grade content acceptance relationship. Then, the connected membrane surface area content is used as the risk write-back range to generate the risk write-back range.

[0138] The aging level content and change-related content within the risk write-back scope are verified for correspondence. The aging level content within the risk write-back scope is then linked to the change-related content in the corresponding area, ensuring that the aging level content maintains a correspondence with the change-related content. Aging level content that continuously corresponds to change-related content is retained within the corresponding write-back scope, allowing it to continue participating in the risk write-back scope. Aging level content with a broken correspondence is separated from the corresponding write-back scope, preventing it from remaining within the scope. Finally, the risk write-back scope after verification is used as the verified risk write-back content, generating the verified risk write-back content.

[0139] After verification, the risk write-back content is connected to the image content in the membrane area. The aging candidate region in the verified risk write-back content is connected to the image content in the same membrane area, so that the aging candidate region can correspond to the specific image content. The aging level content is connected to the image content in the same membrane area, so that the aging level content can be consistent with the corresponding image content. The region position is connected to the image content in the same membrane area, so that the image content can take over the corresponding region position. Then, the aging candidate region, aging level content, region position and corresponding image content are kept in correspondence in the same membrane area to generate image-associated write-back content.

[0140] Image-associated write-back content is embedded into the membrane area image data according to its regional location and connectivity. The image-associated write-back content is then connected to the membrane area content at the same location in the membrane area image data, ensuring that the image-associated write-back content maintains positional consistency with the membrane area content. The image-associated write-back content is then connected to adjacent membrane area content along the regional connectivity, ensuring that the risk write-back range maintains positional consistency with the corresponding membrane area content. Finally, the image-associated write-back content that has been embedded and maintains positional consistency is used as the membrane risk write-back content to generate the membrane risk write-back content.

[0141] The aging candidate region, aging level content, region location, region connection relationship, change-related content, and corresponding image content in the membrane risk write-back content are encapsulated. The aging candidate region is connected to the corresponding region location, so that the aging candidate region can maintain positional correspondence in the service membrane aging test results; the aging level content is connected to the corresponding aging candidate region, so that the aging level content can correspond to the specific aging candidate region; the region connection relationship is connected to the corresponding region location, so that the connection between adjacent region locations can be preserved; the change-related content is connected to the corresponding aging level content and aging candidate region, so that the change-related content can maintain correspondence with the aging level content; and the corresponding image content is connected to the same membrane region, so that the service membrane aging test results can simultaneously inherit each corresponding content in the membrane risk write-back content and generate the service membrane aging test results.

[0142] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An image recognition-based aging detection system for in-service membrane materials, characterized in that: include: The in-service image acquisition module receives in-service membrane material image data, performs membrane surface area processing on the in-service membrane material image data, and generates membrane surface area image data. The image normalization processing module receives image data of the membrane surface area, performs service image normalization processing on the membrane surface area image data, and generates service membrane material normalized image data. The aging appearance feature construction module receives normalized image data of the in-service membrane material, performs aging appearance feature construction processing on the normalized image data of the in-service membrane material, and generates aging appearance feature data. The historical change association module receives aging appearance feature data and historical membrane surface image data, performs historical change association processing on the aging appearance feature data and historical membrane surface image data, and generates aging change association data. The aging candidate region identification module receives aging appearance feature data and aging change correlation data, performs aging candidate region identification processing based on the Chan-Vese segmentation algorithm on the aging appearance feature data and aging change correlation data, and generates aging candidate region data. The aging level assessment module receives aging candidate area data and aging change correlation data, performs aging level assessment processing on the aging candidate area data and aging change correlation data, and generates aging level assessment data. The membrane risk write-back module receives aging level assessment data and membrane area image data, performs membrane risk write-back processing on the aging level assessment data and membrane area image data, and generates aging test results of the in-service membrane material. The steps in generating aging appearance feature data by the aging appearance feature construction module include: The aging appearance feature construction module receives normalized image data of the service membrane material and arranges the membrane surface area content, region position and region connection relationship in the normalized image data of the service membrane material according to the region position, so that the content of each membrane surface area forms a continuous connection relationship within the same membrane surface area. The image color expressions in each membrane area are aggregated according to the region location, and the image color expressions within the same membrane area and between adjacent membrane areas are compared according to the region connection relationship. The membrane area content where the color expression shows continuous differences is marked as color change content. The image brightness and color representations in each membrane area are combined according to their location, and the corresponding content after combination is continuously compared according to the connection relationship of the regions. The corresponding content that appears repeatedly in adjacent regions and is arranged in a continuous direction is marked as texture change content. The corresponding content that extends continuously in the texture change content is connected to the color change content in terms of position, and the extension relationship between adjacent film areas is verified along the region connection relationship. The corresponding content that crosses the position of adjacent areas and maintains continuous extension is marked as crack continuity content. The continuous edge content in the membrane area is compared with the region connection relationship, and the corresponding content that is broken, misaligned or missing in the continuous edge content is marked as boundary damaged content. Connect the color change content, texture change content, crack continuity content, and boundary damage content according to their regional positions, so that each corresponding content is consistent with the same film surface area. The color change content, texture change content, crack continuity content, and boundary damage content of the completed location connection are encapsulated according to the film surface area content to generate aging appearance feature data. The steps by which the aging candidate region identification module generates aging candidate region data include: The aging candidate region identification module receives aging appearance feature data and aging change correlation data, and connects and arranges the color change content, texture change content, crack continuity content, boundary damage content, and change correlation content corresponding to the same film surface area according to the region location. The corresponding content after arrangement is imported into the level set initialization process of the Chan-Vese segmentation algorithm to form an initial level set outline within the region location, and to maintain the positional correspondence between the initial level set outline and the aging appearance feature data and the aging change correlation data. Divide the inner and outer regions along the initial horizontal set outline. Collect the color change content, texture change content, crack continuity content and boundary damage content corresponding to the inner region. Collect the color change content, texture change content, crack continuity content and boundary damage content corresponding to the outer region. Match the corresponding content in the inner and outer regions with the change-related content of the same membrane region, write the content of regions with continuous change relationship into the inner evolution direction of the horizontal set, and write the content of regions without continuous change relationship into the outer exclusion direction of the horizontal set. The color change content, texture change content, crack continuity content, and boundary damage content in the evolution direction inside the horizontal set are uniformly merged so that the appearance content with continuous change relationship participates in the attraction of aging region. The corresponding content in the exclusion direction outside the horizontal set is input into the local Reed-Xiaoli anomaly detection algorithm to form a local neighborhood around the region. The local distribution relationship between the color change content, texture change content and boundary damage content in the local neighborhood is calculated. The deviation degree between the corresponding content of the position to be judged in the exclusion direction outside the horizontal set and the local distribution relationship is calculated to form an anomaly deviation judgment result. The local appearance content that lacks continuous change relationship and forms an anomaly deviation judgment result is regarded as interference anomaly content. The aging region attracts and interferes with abnormal content, which is written into the region evolution process of the Chan-Vese segmentation algorithm. This causes the level set contour to advance towards regions with continuous aging change relationships and exit from regions corresponding to the interference abnormal content. The contours of the level set are continuously evolved according to the Chan-Vese segmentation algorithm, and the contour movement content and change-related content in adjacent evolution sequences are verified accordingly. Retain the outline content that continuously corresponds to the content associated with the change, and remove the outline content whose correspondence is interrupted, so that the outline of the level set converges to the aging boundary position. The continuous connected film surface regions within the converged aging boundary are merged, and the merged region content is connected with the corresponding color change content, texture change content, crack continuity content, boundary damage content, and change association content to generate aging candidate region data.

2. The service membrane material aging detection system based on image recognition according to claim 1, characterized in that: The steps by which the in-service image acquisition module generates image data of the membrane surface area include: The service image acquisition module receives service membrane material image data, arranges the image content in the service membrane material image data according to the acquisition order, and generates the sequentially arranged image content; The image content after sequential arrangement is divided into the membrane surface area content and the background area content, and the membrane surface area content is retained; Connect the continuous edge content and the continuous internal content in the membrane area according to the membrane area to generate continuous membrane area content; The content of a continuous membrane surface region is divided into multiple membrane surface regions, and each membrane surface region is used as a region location. Preserve the positional relationship of each membrane region within the content of a continuous membrane region, and connect membrane regions that are adjacent in position to generate region connection relationships; The membrane surface area content, location, and connectivity of each membrane surface area are connected to generate membrane surface area image data.

3. The service membrane aging detection system based on image recognition according to claim 1, characterized in that: The steps of the image normalization processing module in generating normalized image data of in-service membrane materials include: The image normalization processing module receives the image data of the membrane area and arranges the content, location and connection relationship of each membrane area according to the location, so that the content of each membrane area forms a continuous arrangement relationship in terms of location. The image brightness and darkness representations in each membrane area are collected according to the region location, and the image brightness and darkness representations of adjacent membrane areas are compared according to the region connection relationship. The brightness and darkness differences obtained from the comparison are applied to the corresponding membrane area content, so that the content of adjacent membrane areas forms a continuous brightness and darkness transition along the region connection relationship. The image color representations in the membrane area content after the completion of continuous light and dark transitions are aggregated according to the region position, and the image color representations of adjacent membrane area content are compared according to the region connection relationship. The color differences obtained from the comparison are applied to the corresponding membrane area content, so that the content of adjacent membrane area content forms a continuous color transition along the region connection relationship. Connect the local light and dark abrupt changes in the content of the membrane area after the continuous color transition with the continuous light and dark transition of the adjacent membrane area content, so that the local light and dark abrupt changes are incorporated into the light and dark continuity of the corresponding area position. After completing the connection of light and dark continuity, the content of the membrane area is aligned according to the region position, so that the content of the membrane area corresponding to the same membrane area falls into the same region position, and the correspondence between the aligned membrane area content and the region connection relationship is maintained. Connect the contents of adjacent membrane areas with continuous regional connectivity to the boundary, so that the contents of adjacent membrane areas are included in the same membrane area. The content, location, and connection relationships of the membrane area after completing continuous light and dark transitions, continuous color transitions, continuous light and dark relationship connections, area position alignment, and boundary connection are encapsulated to generate normalized image data of the membrane material in service.

4. The service membrane aging detection system based on image recognition according to claim 1, characterized in that: The steps for the historical change correlation module to generate aging change correlation data include: The historical change association module receives aging appearance feature data and historical film surface image data, and arranges the color change content, texture change content, crack continuity content and boundary damage content in the aging appearance feature data according to the regional position, so that each corresponding content is consistent with the position of the film surface area. Historical membrane surface image data is mapped to aging appearance feature data according to regional location, so that the image content corresponding to the same membrane surface area in the historical membrane surface image data falls into the same regional location, and the image content in the same regional location has corresponding image color expression, image brightness expression, edge continuity content and regional connection relationship. By comparing the color representation and color change content of the same membrane area in the historical membrane image data, the continuity relationship of color change of the same membrane area in the historical membrane image data and the aging appearance feature data is determined. By comparing the image brightness and color representation with the texture change content corresponding to the same membrane area in the historical membrane image data, the continuity relationship of texture change between the same membrane area in the historical membrane image data and the aging appearance feature data is determined. The texture variation continuity relationship and the crack continuity content are connected according to the region connection relationship to determine the crack variation continuity relationship that spans adjacent regions and maintains extension between historical film surface image data and aging appearance feature data. By comparing the continuous edge content and boundary damage content corresponding to the same membrane area in the historical membrane image data according to the regional connection relationship, the boundary damage continuity relationship of the same membrane area between the historical membrane image data and the aging appearance feature data is determined. The relationships of color change continuity, texture change continuity, crack change continuity, and boundary damage continuity are connected according to the region location to form the change association content corresponding to the same film area; The change-related content is positionally linked with the color change content, texture change content, crack continuity content, and boundary damage content in the aging appearance feature data to generate aging change-related data.

5. The service membrane material aging detection system based on image recognition according to claim 1, characterized in that: The steps of the local Reed-Xiaoli anomaly detection algorithm are as follows: Arrange the corresponding content in the outer exclusion direction of the horizontal set according to the region position, so that the color change content, texture change content, boundary damage content and change-related content maintain the correspondence in the same region position; A local neighborhood is defined around the position to be judged in the exclusion direction outside the horizontal set, and the color change content, texture change content, and boundary damage content adjacent to the position to be judged within the local neighborhood are collected; The color change content, texture change content, and boundary damage content gathered in the local neighborhood are combined according to the same region location, and the local distribution relationship between the color change content, texture change content, and boundary damage content in the local neighborhood is calculated. The Reed-Xiaoli deviation of the content corresponding to the position to be judged in the direction of exclusion outside the horizontal set is calculated with respect to the local distribution relationship, thus forming the abnormal deviation judgment result of the content corresponding to the position to be judged relative to the local neighborhood; The continuity relationship between the abnormal deviation judgment result and the change-related content corresponding to the position to be judged is checked to determine whether the content corresponding to the position to be judged is missing a continuous change relationship; Local appearance content that lacks continuous change relationships and forms abnormal deviations from the discrimination results is regarded as interfering abnormal content.

6. The service membrane aging detection system based on image recognition according to claim 1, characterized in that: The steps involved in generating aging level assessment data by the aging level assessment module include: The aging level assessment module receives aging candidate area data and aging change correlation data. It connects the aging candidate areas in the aging candidate area data with the change correlation content corresponding to the same membrane area in the aging change correlation data according to the area location, and generates a level assessment object. The continuity of color changes in the assessment objects is judged, and the extension range of color changes in the same aging candidate area, the connection between adjacent areas, and the corresponding changes are compared to form the basis for color change level. The coarsening degree of texture changes in the grade assessment object is judged, and the recurrence range, arrangement direction continuity and corresponding change-related content of texture changes in the same aging candidate area are compared to form the basis for texture change grade. The extent of expansion of the crack continuity content in the grade assessment object is judged. The extension relationship of the crack continuity content across the adjacent area, the positional connection relationship between the crack continuity content and the color change content and texture change content, and the corresponding change-related content are compared to form the basis for crack continuity grade. The degree of damage to the boundary damage content in the grade assessment object is judged. The extent of the breakage, misalignment or missing content at the boundary of the aging candidate area and the corresponding changes are compared to form the basis for the boundary damage grade. The color change level criteria, texture change level criteria, crack continuity level criteria, and boundary damage level criteria are connected according to the region location, and the level criteria that are continuously associated with the change content are retained in the same aging candidate region to generate the region level criteria. Based on the combination of color change, texture change, crack continuity and boundary damage in the regional grading criteria, each aging candidate region is graded and aging grade content is generated. The aging level information is linked with the corresponding aging candidate areas, area locations, and change-related information to generate aging level assessment data.

7. The service membrane material aging detection system based on image recognition according to claim 1, characterized in that: The steps for the membrane risk write-back module to generate aging test results for in-service membrane materials include: The membrane risk write-back module receives aging level assessment data and membrane area image data. It arranges the aging level content, aging candidate area, area location and change association content in the aging level assessment data according to the area location, and matches them with the membrane area content and area connection relationship in the membrane area image data to generate write-back positioning content. Write the aging level content from the write-back positioning content into the film surface area content corresponding to the same area location, so that each aging candidate area has the corresponding aging level content in the film surface area image data, and generate the level write-back content. The aging candidate area in the grade write-back content is connected to the membrane area with the adjacent area connection relationship. The membrane area content that crosses the adjacent area position and has continuous aging grade content is connected into the same write-back range to generate the risk write-back range. The aging level content and change-related content in the risk write-back scope are verified to correspond. The aging level content that continuously corresponds to the change-related content is retained in the corresponding write-back scope, and the aging level content with interrupted correspondence is separated from the corresponding write-back scope to generate verified risk write-back content. The verified risk write-back content is linked with the image content in the membrane area content, so that the aging candidate area, aging level content, area location and corresponding image content maintain a corresponding relationship under the same membrane area, and image-related write-back content is generated. The image-associated write-back content is embedded into the membrane area image data according to the regional location and regional connection relationship, so that the risk write-back range and the corresponding membrane area content are kept in the same position, and membrane risk write-back content is generated. The aging candidate area, aging level, area location, area connection relationship, change-related content and corresponding image content in the membrane risk write-back content are encapsulated to generate the aging test results of the membrane material in service.

Images