Offshore wind turbine pile foundation disease identification and detection method and system based on image processing

Abstract

The application discloses a kind of offshore wind turbine pile foundation disease identification detection method and system based on image processing, it is related to pile foundation structure health monitoring technical field;Acquisition inspection image and pile foundation structure parameter and historical data, and pixel point is mapped to pile foundation cylindrical coordinate system;Based on image quality and resolution calculation image credibility, image is processed in stages;According to different functional subarea of tidal level and mud surface line to pile foundation, and configure different disease weight;In each subarea, corrosion, crack, coating peeling and marine organism attachment and other diseases are identified and the severity index is calculated;Further, the disease area ratio change is analyzed by historical image registration, and the disease development trend value is obtained;Finally, the overall risk index is generated by fusing the severity index and the development trend, and the inspection report is output, which can improve the accuracy and stability of disease identification in complex marine environment, and realize the transformation of pile foundation disease from static identification to dynamic risk assessment.

Description

Technical Field

[0001] This invention relates to the field of pile foundation health monitoring technology, and more specifically, to a method and system for identifying and detecting defects in offshore wind turbine pile foundations based on image processing. Background Technology

[0002] As the service life of offshore wind turbines increases, the foundations of offshore wind turbines are subjected to a complex environment of salt spray, tides, splash, seawater immersion and mud erosion for a long time. Their surfaces are prone to corrosion, cracks, coating peeling and marine organism attachment.

[0003] Current pile foundation inspection methods typically rely on drones, ROV underwater robots, or manual verification to collect images, which are then used by inspectors to make judgments based on visible defects in the images. While this method can detect some obvious defects, it still has the following shortcomings in practical applications: On the one hand, marine inspection images are easily affected by factors such as water turbidity, reflection, obstruction, changes in shooting angle, and inconsistent scale resolution. If the disease level is judged directly based on the image results, it is easy to misjudge blurry areas, areas obscured by marine organisms, or local reflective areas as real diseases. On the other hand, images from different inspection batches often come from different shooting angles and coverage areas. If the pixel positions in the images cannot be uniformly mapped to the coordinates of the pile foundation surface, it is difficult to accurately align historical images with current images. This leads to the same corrosion patch, crack, or coating peeling area being repeatedly counted or incorrectly compared at different times. In addition, the atmospheric zone, splash zone, fully immersed zone, and mud-surface adjacent zone of the pile foundation are in different environments, and the types of diseases and their structural impact are also different. If the existing methods use a unified identification standard and a unified weight for evaluation, it is difficult to accurately reflect the real risks of different functional zones. In addition, traditional disease identification and detection methods mostly focus on describing the currently visible disease area, lacking comprehensive calculations of changes in the proportion of disease area, the expansion trend per unit time, and the importance of zoning structure, making it difficult to form quantifiable and traceable overall risk results.

[0004] To address this, a method and system for identifying and detecting defects in offshore wind turbine pile foundations based on image processing has been developed. Summary of the Invention

[0005] To overcome the above-mentioned deficiencies of the prior art, embodiments of the present invention provide a method and system for identifying and detecting defects in offshore wind turbine pile foundations based on image processing.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The image processing-based method for identifying and detecting defects in offshore wind turbine pile foundations is as follows: S1: Obtain the inspection images of offshore wind turbine pile foundations and their corresponding pile foundation structural parameters and historical inspection parameters, and map the pixel positions in the inspection images to the pile foundation cylindrical coordinate system; S2: Calculate the image credibility for each inspection image. The image credibility is calculated based on the image quality score and scale resolution score. Determine whether the inspection image is included in the calculation of the deterministic disease level based on the image credibility. S3: After entering the deterministic disease level calculation, the pile foundation surface is divided into atmospheric zone, splash tidal zone, full immersion zone and mud surface adjacent zone according to the axial height of the pile foundation, tidal range and mud surface line position, and different disease characteristic weights are assigned to different functional zones. S4: Identify corrosion, cracks, coating peeling and marine organism attachment in each functional zone to obtain image feature parameters of various diseases. Use the image feature parameters combined with the disease feature weights of each functional zone to output the disease severity index of each functional zone. S5: Register the current inspection image and the historical inspection image to the same pile foundation cylindrical coordinate system, calculate the change in the area ratio of various types of defects, and obtain the defect development trend value of each functional zone; S6: Convert the disease severity index and the disease development trend value into the overall risk index of the pile foundation, and combine them with the image credibility to fill the pre-edited report template to generate a pile foundation inspection report.

[0007] Specifically, image credibility Retrieval logic; ; in Indicates the image quality score. Indicates the scale resolution score. , These are the weighting coefficients, and their sum is one. Image quality score It is obtained by weighting the sharpness score, brightness score, contrast score, and occlusion score; The sharpness score is obtained by normalizing the Laplacian variance of the grayscale image of the inspection image; The brightness score is determined by the deviation of the average brightness of the target area from the preset target brightness median value; The contrast ratio is obtained by normalizing the standard deviation of the target area brightness relative to the preset minimum effective contrast ratio and the upper limit of the contrast ratio; The occlusion score is determined by the ratio of the pixel area of ​​the occluded region to the total area of ​​the target recognition region; The scale resolution score is obtained by normalizing the actual resolution of the target area relative to the preset high-precision recognition resolution and the upper limit of acceptable resolution. When the image confidence level is lower than the first confidence threshold, the output image is not suitable for direct labeling; when the image confidence level is between the first and second confidence thresholds, manual review of the labeling is triggered; when the image confidence level is higher than the second confidence threshold, the calculation of the deterministic disease level is initiated.

[0008] Specifically, the pile foundation surface is divided into an atmospheric zone, a splash and tidal zone, a fully submerged zone, and a mud-surface adjacent zone; Obtain the axial height of the pile foundation Historical high tide level Historical low Mud surface height ; Mud surface adjacent area Defined as: ;in The width of the adjacent area of ​​the mud surface is half its width; splash zone Defined as: ;in This is a pre-set tidal range allowance. To allow for the pre-set splashing of waves; Atmospheric region Defined as: ; Full immersion area Defined as: ; The disease weights corresponding to each functional zone can be set as follows: i = 1, 2, 3, or 4; where These respectively represent corrosion, cracks, coating peeling, and marine organism attachment in the zones. Weights within.

[0009] Specifically, corrosion, cracks, coating peeling, and marine organism attachment are identified in each functional zone, and a disease severity index is output. Let the pixel set of the i-th functional partition be denoted as Let the identification mask for the m-th type of disease be denoted as ,in ; These respectively represent corrosion, cracks, coating peeling, and marine organism attachment; For pixel p, if it belongs to the m-th type of defect, then =1, if it does not belong to then =0; The disease area of ​​the m-th type of disease within the i-th functional zone is obtained through... The calculation yielded: in, ap represents the actual area of ​​the m-th type of disease within the i-th functional zone; ap represents the actual area of ​​the pile foundation surface corresponding to pixel p. Based on the actual affected area of ​​various diseases and the effective detection area of ​​the i-th functional zone, calculate the corrosion ratio, crack ratio, coating peeling ratio and marine organism attachment ratio. Multiply the area percentage of each diseased area by the corresponding disease weight in the i-th functional zone and sum them to obtain the disease severity index Si of the i-th functional zone.

[0010] Specifically, the actual area of ​​the pile foundation surface is determined based on the cylindrical coordinates of the pile foundation after the pixel point is mapped; if the cylindrical coordinates of pixel point p are... The coordinate differences between its adjacent pixels in the circumferential and axial directions are respectively and ,but: ; Where D represents the pile diameter; D / 2 represents the pile radius; This represents the angular interval corresponding to pixel p in the circumferential direction; This represents the actual height interval corresponding to pixel p in the axial height direction. This represents the actual arc length of the pixel along the circumference of the pile foundation.

[0011] Specifically, the current inspection image and historical inspection images are registered to the same pile foundation cylindrical coordinate system and the trend value of disease development is calculated; Calculate the disease development trend value of the i-th functional zone. : ; in, This represents the disease development trend value for the i-th functional zone; , , as well as These represent the development scores of various diseases within the i-th functional zone.

[0012] Specifically, the logic for obtaining development scores; The current inspection time is denoted as Tc, the inspection time of the historical inspection image is denoted as Th, and the time interval between the two is calculated and denoted as Tg. Extract the corrosion percentage, crack percentage, coating peeling percentage, and marine organism attachment percentage of the target historical pile foundation corresponding to each functional zone from historical inspection images; Calculate the change in the area percentage of the m-th type of disease within the i-th functional zone: ; and These represent the area percentage of the m-th type of disease in the i-th functional zone at the current inspection time and the area percentage of the m-th type of disease in the i-th functional zone at the historical inspection time, respectively. Calculate the rate of disease spread per unit time: ; The changes in various diseases within the i-th functional zone are normalized to obtain the development score of the m-th disease. : ; in, This represents the development score of the m-th type of disease within the i-th functional zone; This represents the low development rate threshold for the m-th type of disease; This represents the high development rate threshold for the m-th type of disease; clip(·, 0, 1) restricts the calculation result to between 0 and 1.

[0013] Specifically, the disease development trend value and disease severity index are converted into the overall risk index of the pile foundation; Let Si be the disease severity index of the i-th functional zone, combined with the disease development trend value of the i-th functional zone. Using the formula The partition risk value Ri of the i-th functional partition is calculated; where β represents the weighting coefficient of the disease severity index; β represents the weighting coefficient of the disease development trend value; the overall risk index of the pile foundation. .

[0014] The technical effects and advantages of this invention are as follows: This invention introduces an image credibility evaluation mechanism to comprehensively quantify the clarity, brightness, contrast, occlusion degree, and scale resolution of inspection images, and sets graded threshold judgments to prevent low-quality or strongly interfering images from directly entering the disease identification process, thereby reducing misjudgments from the source. Furthermore, it combines semantic segmentation to identify multiple categories of corrosion, cracks, coating peeling, and marine organism attachment, and filters out small-area noise regions through connected components, thereby improving the identification accuracy and making disease detection in complex marine environments more robust and engineering-applicable. This invention maps inspection image pixels to the cylindrical coordinate system of the pile foundation, achieving spatial alignment of images taken from different angles and at different times. This solves the problem of difficulty in comparing defects across time periods in traditional methods. Based on this, the pile foundation is divided into atmospheric zone, splash and tidal zone, fully submerged zone, and mud-surface adjacent zone, and differentiated defect weights are assigned to different environments, making defect assessment more consistent with the actual corrosion mechanism. At the same time, by calculating the change in the area ratio of various defects and the expansion rate per unit time, a defect development trend value is formed, which is further integrated with the defect severity index to obtain the zoning and overall risk index, improving the guiding value of inspection results for operation and maintenance decisions. Attached Figure Description

[0015] Figure 1This is a flowchart of the image processing-based method for identifying and detecting defects in offshore wind turbine pile foundations according to the present invention. Detailed Implementation

[0016] 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.

[0017] Example 1 like Figure 1 As shown, the image processing-based method for identifying and detecting defects in offshore wind turbine pile foundations is as follows: S1, acquire the inspection images of offshore wind turbine pile foundations and their corresponding pile foundation structural parameters and historical inspection parameters, and map the pixel positions in the inspection images to the pile foundation cylindrical coordinate system; In one embodiment, the inspection equipment is a combination of an ROV (Return on Marine Vehicle) and a drone. The ROV is used to acquire images of the underwater area and the area adjacent to the mud surface, while the drone is used to acquire images of the atmospheric area and the splash zone. The inspection image resolution can be 1920×1080 pixels, 2560×1440 pixels, or higher; when used for crack width identification, the actual resolution of the target area should preferably be no less than 1.0 mm / pixel, preferably no less than 0.5 mm / pixel, but this value is not an absolute limitation.

[0018] The kth inspection image is used This indicates that t represents the data acquisition time; For each inspection image Obtain camera intrinsic parameters K and camera extrinsic parameters and target depth When an ROV is equipped with a binocular camera, structured light, laser ruler, or sonar ranging device, the target depth... The target depth is given directly by the corresponding equipment; when only a monocular camera is available, the target depth can be calculated from the pile diameter D, the distance between the two parallel laser spots, or a known scale. Image pixels Ray direction mapped to the camera coordinate system: ; According to depth Obtain the spatial point: ; Then based on camera external parameters Will Transform to the pile foundation coordinate system to obtain spatial points .in This refers to the axial height of the pile foundation. and Located within the horizontal section of the pile foundation.

[0019] For a cylindrical single-pile foundation, the circumferential angle Calculate using the following formula: = ; Therefore, image pixel p is mapped to the coordinate point of the pile foundation cylinder ( , When the pile foundation is not a complete cylinder, such as the inclined brace of the jacket or the local structure of the transition section, the corresponding structural component can be established as a local surface coordinate system, and the location of the defect is still recorded by the two-dimensional unfolding method of "surface circumferential or transverse coordinate + axial coordinate". This coordinate transformation determines whether subsequent historical images can be aligned. Without coordinate mapping of the pile foundation cylinder, the same corrosion patch photographed at different times and angles will be identified as multiple independent diseases, leading to an amplification of the disease's expansion trend.

[0020] The structural parameters of the pile foundation include pile diameter D, wall thickness T, and design mudline height. ; Historical inspection parameters include a set of historical images; S2, calculate the image reliability for each inspection image. Image credibility is based on image quality score. and scale resolution Calculated; ; , These are the weighting coefficients, and their sum is one. In one embodiment, =0.70, =0.30.

[0021] Image quality score is obtained through Calculated; in , , as well as These represent the sharpness score, brightness score, contrast score, and occlusion score, respectively. The sharpness score is calculated using Laplace variance normalization: ; in To measure the Laplacian variance of the grayscale image during inspection; These are preset lower and upper limits for recognizability; for 1920×1080 pixel images, Take 80, A value of 900 is acceptable; this value should be calibrated in conjunction with the camera model and compression parameters.

[0022] Brightness score through Calculated; where The average brightness of the target area in the image. The preset target brightness median value, Preset allowable brightness deviation; In an 8-bit grayscale image 128 is acceptable. 80 is acceptable.

[0023] Comparison score through Calculated; in The standard deviation of the target area's brightness. and The preset minimum effective contrast ratio and maximum contrast ratio; In one embodiment, and Take 15 and 70 respectively; Obstruction score passed Calculated; in and These represent the pixel area of ​​the occluded region and the total area of ​​the target recognition region, respectively. The occlusion area represents the pixel area covered by water droplets, suspended objects, strong reflections, marine life, or lens contamination. It can be determined by combining the high-brightness saturated area, the low-texture mud and sand occlusion area, the lens edge stain template, and the preliminary marine life segmentation results.

[0024] Scale resolution score is calculated using the formula Calculated; in The actual resolution of the target area. and These are the preset resolution required for high-precision recognition and the acceptable upper limit, respectively.

[0025] In one embodiment, and We take 0.5mm / pixel and 3.0mm / pixel respectively.

[0026] When the image confidence level is lower than the preset confidence threshold range, the "image is not suitable for direct judgment" flag is output. When it is within the preset confidence threshold range, the manual review flag is triggered. When it is higher than the threshold range, the calculation of the deterministic disease level is initiated. S3, after entering the deterministic disease level calculation, the pile foundation surface is divided into atmospheric zone, splash tidal zone, full immersion zone and mud surface adjacent zone according to the axial height of the pile foundation, tidal range and mud surface line position, and different disease characteristic weights are assigned to different functional zones. Partitioning logic: Axial height of pile foundation Historical high tide level Historical low Mud surface height ; Mud surface adjacent area Defined as: ;in The width of the adjacent area to the mud surface is half, and in one embodiment, it is taken as 0.5D to 2.0D, where D is the pile diameter. For a single pile with a diameter of 6m, The range is 3m to 12m. The mud surface adjacent area has the highest priority. As long as a pixel falls into this area, it will be treated as a mud surface adjacent area even if it also meets the full immersion area condition.

[0027] splash zone Defined as: ;in The pre-set tidal range extension margin can be taken as 0.2m to 1.0m; The preset splash margin can be taken as 1.0m to 5.0m; for sea areas with larger wave heights, the upper limit range can be adjusted by technical personnel to obtain a higher value; Atmospheric region Defined as: ; Full immersion area Defined as: .

[0028] Functional zoning is not about cropping an image into visually appealing segments, but about altering the rules for interpreting defects. For example, reddish-brown textures in splash zones have higher weighting because this area experiences prolonged wet-dry cycles.

[0029] The disease weights corresponding to each functional zone can be set as follows: i = 1, 2, 3 or 4; in These respectively represent corrosion, cracks, coating peeling, and marine organism attachment in the zones. Weight within; For example, in the splash zone and The weights are relatively high, with the sum of the four weights within each partition being 1.

[0030] S4. In each functional zone, corrosion, cracks, coating peeling, and marine organism attachment are identified to obtain image feature parameters of various diseases. The image feature parameters are combined with the disease weight of each functional zone to output the disease severity index of the i-th functional zone. Specifically: Before identifying the defects, each pixel p is mapped to its surface position in the cylindrical coordinate system of the pile foundation to identify the functional zone and defect category of the target pixel. Let the pixel set of the i-th functional partition be denoted as Let the identification mask for the m-th type of disease be denoted as ; in ; These respectively represent corrosion, cracks, coating peeling, and marine organism attachment; For pixel p, if it belongs to the m-th type of defect, then =1, if it does not belong to then =0; The disease area of ​​the m-th type of disease within the i-th functional zone is obtained through... Calculated; in, Ωi represents the actual area of ​​the m-th type of disease within the i-th functional zone, in m²; Mm(p) represents the set of valid pixels within the i-th functional zone; ap represents the binary recognition result of the m-th type of disease corresponding to pixel p; and ap represents the actual area of ​​the pile foundation surface corresponding to pixel p, in m². Additional explanation: The actual area ap corresponding to pixel p is not a fixed value, but is determined based on the cylindrical coordinates of the pile foundation after the pixel is mapped; if the cylindrical coordinates after pixel p are... The coordinate differences between its adjacent pixels in the circumferential and axial directions are respectively and ,but: ; Where D represents the pile diameter in meters (m); D / 2 represents the pile radius. This represents the angular interval corresponding to pixel p in the circumferential direction, in rad. This represents the actual height interval corresponding to pixel p in the axial height direction, in meters (m). Because... The value 'ap' represents the actual arc length of the pixel along the circumference of the pile foundation. Therefore, 'ap' represents the actual area of ​​the pixel on the cylindrical surface of the pile foundation.

[0031] When an image has been unfolded into a cylindrical unfolded image of equal scale, if the circumferential sampling interval of the unfolded image is Δs and the axial sampling interval is Δz, then the actual area of ​​the pixel can also be simplified to: ap = Δs·Δz. Where Δs represents the actual circumferential arc length corresponding to a pixel or a sampling unit in the unfolded image, in meters (m); Δz represents the actual axial height corresponding to a pixel or a sampling unit in the unfolded image, in meters (m). For example, when Δs = 0.01m and Δz = 0.01m, the actual area corresponding to a pixel is 0.0001m².

[0032] For corrosion diseases, the corrosion area ; in, This represents the erosion area within the i-th functional partition; This indicates an erosion mask. To avoid water droplet reflections, localized blemishes, or image noise being mistaken for erosion, isolated erosion-connected regions with an area less than 0.005 m² can be discarded.

[0033] For crack defects, crack area ; For crack masking; For coating peeling, the area of ​​coating peeling ; As a mask for coating peeling, to avoid misjudging small scratches as peeling, isolated peeling candidate areas with an area of ​​less than 0.01m² can be excluded from the peeling area.

[0034] Regarding marine organism attachment, the area of ​​marine organism attachment. ; For marine organism attachment masking; if marine organism attachments cover areas of corrosion or coating peeling, these areas are included in the marine organism attachment area and marked as "covered area to be cleaned and re-inspected" in the results, without directly determining whether there is corrosion or peeling underneath.

[0035] For the i-th functional zone, after determining the corrosion area, crack area, coating peeling area, and marine organism attachment area, calculate the proportion of different disease areas in the total area of ​​the functional zone, and output the corrosion proportion, crack proportion, coating peeling proportion, and marine organism attachment proportion. The proportions of each group are compared with the disease weights determined by the i-th functional partition. The corresponding weights in the index are multiplied together and then summed to obtain the disease severity index of the i-th functional partition; In one embodiment, the output logic for the four types of disease masks is as follows: The process for determining the etching mask is as follows: Before calculating the corrosion identification score, the image is subjected to white balance correction, brightness equalization and reflective area removal. The corrected image is then converted to Lab color space and HSV color space. The Lab color values ​​of pixel p are recorded as L(p), a(p), b(p), and the HSV color values ​​are recorded as H(p), S(p), V(p). Where L(p) represents the luminance component, a(p) represents the red-green color component, b(p) represents the yellow-blue color component, H(p) represents the hue, S(p) represents the saturation, and V(p) represents the brightness. Set a local neighborhood window Wp around pixel p. The window size can be 31×31 pixels, 51×51 pixels or 101×101 pixels. Calculate the local background color mean based on the pixels in the neighborhood that have not been initially identified as defects, occlusions or strong reflections, and denoted as L0(p), a0(p) and b0(p).

[0036] L0(p), a0(p), and b0(p) are used to represent the reference colors of adjacent complete coatings or normal surfaces within the same functional zone.

[0037] Corrosion color score Calculate according to the following formula: ; ; ; in, This represents the corrosion color score based on the Lab color space; This represents the corrosion color score based on the HSV color space; dH(p) represents the minimum angular distance between the hue H(p) of pixel p and the corrosion reference hue Hcor, in degrees. Hcor can be 25° and is used to represent the reference hue of the reddish-brown to yellowish-brown corrosion area; S(p) represents the saturation of pixel p. clip[,0,1] indicates that the calculation result is limited to between 0 and 1.

[0038] in, Calculate according to the following formula: ; Corrosion texture score Calculate according to the following formula: ; in, The standard deviation of gray values ​​within the neighborhood window Wp of pixel p; The median value of the standard deviation of grayscale values ​​in normal surface areas within the same functional zone; The larger the value, the coarser the texture of the area; corroded areas typically exhibit granular, mottled, or coarse textures, thus the more pronounced the texture fluctuations. The higher.

[0039] Corrosion edge score Calculate according to the following formula: ; ; Where E(p) represents the edge density within the neighborhood window Wp of pixel p; This represents the number of pixels within window Wp whose gradient magnitude is greater than the edge threshold Tg; Indicates the number of effective pixels within window Wp; Tg can be 20, suitable for 8-bit grayscale images; the boundary of corrosion patches is not as sharp as that of cracks, but there are usually local edges and patch outlines, so edge density can be used as an auxiliary judgment quantity; Pitting and Dark Spot Score Calculate according to the following formula: ; Where G(p) represents the gray value of pixel p; This represents the median gray value within the neighborhood window Wp; Indicates the morphology coefficient of the dark spot; When the actual area of ​​the dark connected region to which pixel p belongs is between 0.00001m² and 0.002m², and the circularity of the dark connected region is between 0.20 and 1.00, =1; otherwise =0; Pitting corrosion usually manifests as localized small-area dark pits or dark spots, and is therefore determined by the decrease in brightness and the morphology of small connected regions.

[0040] Corrosion identification score Calculate according to the following formula: ; in, Indicates the corrosion color score; Indicates the corrosion texture score; Indicates the corrosion edge score; This represents the score for pitting dark spots; the sum of all weights is 1.

[0041] when When, pixel p is determined as a candidate pixel for erosion; when At that time, pixel p is not considered as a candidate pixel for erosion. The corrosion identification threshold can be set to 0.55. After connected component analysis, if the actual area of ​​a connected region of a corrosion candidate pixel is less than 0.005 m², the connected region is removed from the corrosion candidate results; after removal, the corrosion mask is obtained. If pixel p belongs to the final erosion region, then =1, otherwise =0.

[0042] Crack mask The determination process is as follows: Cracks are mainly characterized by thin, dark lines, continuous edges, and a large aspect ratio.

[0043] For pixel p within the i-th functional partition, calculate the crack identification score. : ; Where Cdark(p) represents the dark line score, Cline(p) represents the linear structure score, Cconnect(p) represents the connectivity score, and Ccontrast(p) represents the contrast score between the two sides of the crack.

[0044] The dark line score Cdark(p) is calculated according to the following formula: ; Where G0(p) represents the median gray value of the normal background within the neighborhood window of pixel p, and G(p) represents the gray value of pixel p. Cracks are usually darker than the surrounding coating or metal surface, so the larger G0(p)-G(p) is, the higher Cdark(p) will be.

[0045] The linear structure score Cline(p) is determined by the multi-directional linear filter response: ; in, This represents the maximum linear filtering response of pixel p in the directions of 0°, 15°, 30°, 45°, 60°, 75°, 90°, 105°, 120°, 135°, 150°, and 165°. This indicates the upper limit of the preset linear response, which can be obtained by calibration using clear crack samples.

[0046] A higher linear filter response indicates that the pixel is more likely to be located on a fine crack structure.

[0047] The connectivity score Cconnect(p) is calculated according to the following formula: ; in, This represents the aspect ratio of the candidate dark line connected region to which pixel p belongs; the thinner and longer the candidate region, the better. The larger, The higher.

[0048] Contrast ratio on both sides of the crack Calculate according to the following formula: ; in, This represents the average grayscale difference obtained by sampling on both sides of the candidate crack skeleton line along its normal direction. The clearer the edges of the crack, the better. The higher.

[0049] when ≥ At that time, pixel p is determined as a candidate pixel for crack. A value of 0.60 can be used. Connectivity analysis and skeletonization are performed on the candidate pixels for cracks, retaining only candidate regions that simultaneously meet the following conditions: actual length not less than 0.05m; aspect ratio not less than 5; average width not greater than 20mm. Candidate regions meeting these conditions form a crack mask. If pixel p belongs to a crack region, then Mcrk(p) = 1; otherwise, Mcrk(p) = 0.

[0050] Coating peeling mask The process of determining is as follows.

[0051] Coating peeling typically manifests as flaky color abrupt changes, exposure of the substrate or primer, relatively continuous boundaries, and a large affected area. The coating peeling detection score Rcoat(p) is calculated for pixel p: ; in, Indicates the color difference score. Indicates the substrate or primer exposure score. Indicates the score of the peeling boundary. This indicates the score for a patchy region.

[0052] Color difference score Calculate according to the following formula: ; ; in, The difference in Lab color between pixel p and the local normal coating background is shown. The larger the value, the more likely the pixel is to belong to an area with coating damage, exposed primer, or exposed substrate.

[0053] Substrate or primer exposure score Calculate according to the following formula: ; in, Indicates the exposure score of the metal substrate. This indicates the primer exposure score. It can be determined based on the characteristics of low-saturation, grayish-white, or dark gray metallic colors; The color can be determined based on the primer color sample for the engineering coating. If a primer color sample is unavailable, Set to 0, and judge coating peeling based solely on color difference and boundary features.

[0054] Peeling Boundary Score Calculate according to the following formula: ; in, This indicates the degree of closure of the candidate region boundary to which pixel p belongs, with a value ranging from 0 to 1; flaky peeling regions typically have relatively continuous boundaries, therefore... The larger, The higher.

[0055] Score for patchy regions Calculate according to the following formula: ; in, This represents the actual area of ​​the candidate region to which pixel p belongs, in m²; the closer the area is to the sheet-like peeling feature, the better. The higher.

[0056] when At that time, pixel p is determined as a candidate pixel for coating peeling. A value of 0.55 can be used. Connectivity analysis is performed on the candidate regions for coating peeling to remove isolated regions with an actual area less than 0.01 m², and to remove elongated scratch regions with an aspect ratio greater than 20 and a width less than 5 mm, thus obtaining the coating peeling mask Mcoat(p). If pixel p belongs to the coating peeling region, Mcoat(p) = 1; otherwise, Mcoat(p) = 0.

[0057] Marine organisms attaching to the mask The process of determining is as follows.

[0058] Marine organism attachment typically manifests as granular, clustered, strip-like, or raised coverage areas; the marine organism attachment identification score Rbio(p) is calculated for pixel p: ; in, Indicates the texture score of marine organisms. This indicates the morphological score of marine organisms. Indicates the color score of marine life. This indicates the score for the height of the protrusion.

[0059] Marine life texture score Calculate according to the following formula: ; in, This represents the standard deviation of gray levels within the neighborhood window of pixel p. This represents the median value of the grayscale standard deviation for normal surface areas; areas with marine organism attachment typically exhibit strong granular or clustered textures, therefore, the greater the texture fluctuation, The higher.

[0060] Marine organism morphology score Calculate according to the following formula: ; in, The clustering score represents the density of small-particle targets within a candidate region. The boundary roughness score represents the degree of irregularity at the candidate region boundary. Marine organism attachments are generally not regular geometric shapes; therefore, the higher the clustering degree and the roughness of the boundary, the better. The higher.

[0061] Marine life color score Determined based on the common color range of marine organism attachments: ; in, This indicates the score for white or grayish-white attachments on shellfish or barnacles; This indicates the score for green algae deposits; Indicates the score for dark-colored clusters of attachments; Each color score can be determined by the range of hue, saturation, and lightness in the HSV color space. For example, green algae can correspond to the region where H(p) is between 80° and 160° and S(p) is not less than 0.20; white shellfish can correspond to the region where S(p) is not higher than 0.25 and V(p) is not lower than 0.55.

[0062] Protrusion height score Calculate according to the following formula: ; Where h(p) represents the convex height of pixel p relative to the fitted surface of the pile foundation, in meters (m). h(p) can be obtained from binocular camera, structured light, laser ruler, or sonar ranging data. When depth data is unavailable, Not included in the calculation, Rbio(p) is recalculated using the weights of the first three terms; when At that time, pixel p is identified as a candidate pixel for marine organism attachment. A value of 0.55 can be used. Connectivity analysis is performed on the candidate regions, retaining attachment areas with an actual area not less than 0.002 m², resulting in the marine organism attachment mask Mbio(p). If pixel p belongs to a marine organism attachment area, then Mbio(p) = 1; otherwise, Mbio(p) = 0.

[0063] The conflict handling between different disease masks is as follows: Since the same pixel may meet multiple disease candidate conditions at the same time, it is necessary to perform category priority processing to avoid duplicate area statistics; the priority order is set as: marine organism attachment, cracks, coating peeling, corrosion; When the same pixel is simultaneously identified as marine organism attachment and corrosion or coating peeling, it is preferentially classified into the marine organism attachment mask, and a "masked and awaiting cleaning and re-inspection" mark is added to the area; when the same pixel is simultaneously identified as crack and corrosion, it is preferentially classified into the crack mask, and a corrosion background mark is recorded in the crack neighborhood; when the same pixel is simultaneously identified as coating peeling and corrosion, if the area has sheet-like peeling boundaries and substrate exposure features, it is preferentially classified into the coating peeling mask, and the corrosion status is recorded as an additional attribute of the peeling area.

[0064] In some embodiments, a multi-class semantic segmentation model is used to output various disease masks: The output can be from a single multi-class segmentation network or from multiple independent detectors. The segmentation network can be U-Net, DeepLab, Mask R-CNN, or a lightweight Transformer segmentation network. Training samples should be labeled with at least the disease category, disease boundary, pile foundation functional zoning, and image quality level.

[0065] S5, register the current inspection image and the historical inspection image to the same pile foundation cylindrical coordinate system, calculate the change in the area ratio of various types of defects, and obtain the defect development trend value; Specifically: The current inspection time is denoted as Tc, the inspection time of the historical inspection image is denoted as Th, and the time interval between the two is calculated and denoted as Tg. To avoid misjudgment of area changes caused by different inspection perspectives, both the current inspection image and the historical inspection image are first expanded to a two-dimensional surface coordinate plane under the cylindrical coordinate system of the pile foundation. The two-dimensional surface coordinate plane consists of circumferential coordinates and axial height.

[0066] The set of valid pixels of the i-th functional zone in the current inspection image With the set of valid pixels in historical inspection images Perform matching; denote the effective overlapping region of the two regions under the same cylindrical coordinate grid as... ; in, This indicates the same pile foundation surface area that can be effectively covered by both the current inspection image and historical inspection images; subsequent changes in area only occur when... The calculation is performed within the current image or within a region that is only visible in the current or historical images. The reason for this is that if the current image captures an additional eroded area, but the historical image does not cover that area, directly comparing the total area would misjudge "changes in the shooting range" as "disease expansion".

[0067] Extract the corrosion percentage, crack percentage, coating peeling percentage, and marine organism attachment percentage of the target historical pile foundation corresponding to each functional zone from historical inspection images; Calculate the change in the area percentage of the m-th type of disease within the i-th functional zone: ; and These represent the area percentage of type m disease in the i-th functional zone at the current inspection time and the area percentage of type m disease in the i-th functional zone at the historical inspection time, respectively.

[0068] in This represents the change in the area percentage of the m-th type of disease within the i-th functional zone relative to the historical inspection time. When A value greater than 0 indicates an increase in the proportion of this type of disease area; when When the value is close to 0, it indicates that the disease is basically stable; when When the value is less than 0, it indicates that the proportion of this type of disease area has decreased, which may correspond to a reduction in marine life after coating repair and cleaning or changes in image occlusion. It is necessary to make a judgment based on maintenance records and cleaning records.

[0069] Calculate the rate of disease spread per unit time: ; Additional explanation, in which, This represents the rate of change of the area percentage of the m-th type of disease within the i-th functional zone per unit time.

[0070] The changes in various diseases within the i-th functional zone are normalized to obtain the development score of the m-th disease. : ; in, This represents the development score of the m-th type of disease within the i-th functional zone; This represents the preset low development rate threshold for the m-th type of disease; This represents the preset high development rate threshold for the m-th type of disease; clip(·, 0, 1) restricts the calculation result to between 0 and 1. Below If so, it is considered that the development trend of this type of disease is relatively weak; if Higher than If so, it is considered that the development trend of this type of disease is obvious.

[0071] After obtaining the development scores of various diseases, the disease development trend value of the i-th functional zone is calculated by combining the disease weights corresponding to the i-th functional zone. : ; in, This represents the disease development trend value for the i-th functional zone; , , as well as These represent the development scores of various diseases within the i-th functional zone.

[0072] S6, after converting the disease development trend value and disease severity index into the overall risk index of the pile foundation, combines the image credibility and fills it into the pre-edited report template to generate an inspection report; Specifically: Let Si be the disease severity index of the i-th functional zone, combined with the disease development trend value of the i-th functional zone. Using the formula The partition risk value Ri of the i-th functional partition is calculated; where β represents the weighting coefficient of the disease severity index; β represents the weighting coefficient of the disease development trend value. In one embodiment, and The sum is 1. Take 0.70, The value is set to 0.30 because the current severity of the damage directly reflects the current condition of the pile foundation and should be used as the primary basis for judgment; the trend of damage development reflects the subsequent rate of deterioration and should be used as a basis for early warning.

[0073] Based on the structural importance of each functional area, the structural weights of each functional area are preset; Additional notes: The overall weight of each functional zone can be determined based on its structural importance and effective detection area. : ; in, This represents the overall weight of the i-th functional partition in the overall risk calculation; This represents the pre-defined structural importance coefficient for the i-th functional partition; represents the detection area of ​​the i-th functional zone participating in the risk calculation during the current inspection; j represents the functional zone number participating in the overall risk calculation.

[0074] In one embodiment, the structural importance coefficients of the atmospheric zone, the splash zone, the fully submerged zone, and the mud-surface adjacent zone are... These can be set to 0.80, 1.30, 1.00, and 1.40 respectively.

[0075] Overall risk index of pile foundation .

[0076] Example 2 Based on the image processing-based method for identifying and detecting defects in offshore wind turbine pile foundations provided in Embodiment 1 of this application, Embodiment 2 of this application proposes an image processing-based system for identifying and detecting defects in offshore wind turbine pile foundations. Embodiment 2 is merely a preferred embodiment of Embodiment 1, and its implementation will not affect the individual implementation of Embodiment 1.

[0077] Specifically, the image processing-based offshore wind turbine pile foundation defect identification and detection system provided in Embodiment 2 of this application includes: The image acquisition and coordinate mapping module is used to acquire inspection images, pile structure parameters and historical inspection parameters of offshore wind turbine pile foundations, and map the pixel positions in the inspection images to the cylindrical coordinate system of the pile foundation. The image credibility calculation module is used to calculate the image credibility based on the image quality score and scale resolution score of the inspection image, and output the image credibility as an unsuitable for direct judgment, manual review, or a deterministic disease level calculation instruction. The functional zoning and weight configuration module is used to divide the pile foundation surface into atmospheric zone, splash tidal zone, fully immersed zone and mud surface adjacent zone according to the pile foundation axial height, tidal range and mud surface line position, and to configure the disease weights of corrosion, cracking, coating peeling and marine organism attachment for different functional zones. The disease identification and severity calculation module is used to identify corrosion, cracks, coating peeling and marine organism attachment in each functional zone, generate corresponding disease masks, calculate the area ratio of each type of disease, and output the disease severity index of each functional zone in combination with the disease weight. The historical registration and trend analysis module is used to register the current inspection image and the historical inspection image to the same pile foundation cylindrical coordinate system, calculate the change in the area ratio of various diseases in the overlapping effective area, and output the disease development trend value of each functional zone. The risk calculation and report generation module is used to calculate the overall risk index of the pile foundation based on the severity index, trend value, structural importance coefficient and effective inspection area of ​​each functional zone, and generate a pile foundation inspection report in combination with image credibility.

[0078] The above formulas are all dimensionless calculations. Dimensionless calculations can be performed using various methods such as standardization, which will not be elaborated here. The formulas are derived from software simulations based on a large amount of collected data, and the preset parameters in the formulas can be set by those skilled in the art according to the actual situation.

[0079] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, ATA hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state ATA hard disk.

[0080] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0081] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0082] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0083] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0084] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0085] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable ATA hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0086] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for identifying and detecting defects in offshore wind turbine pile foundations based on image processing, characterized in that, include: S1: Obtain the inspection images of offshore wind turbine pile foundations and their corresponding pile foundation structural parameters and historical inspection parameters, and map the pixel positions in the inspection images to the pile foundation cylindrical coordinate system; S2: Calculate the image credibility for each inspection image. The image credibility is calculated based on the image quality score and scale resolution score. Determine whether the inspection image is included in the calculation of the deterministic disease level based on the image credibility. S3: After entering the deterministic disease level calculation, the pile foundation surface is divided into atmospheric zone, splash tidal zone, full immersion zone and mud surface adjacent zone according to the axial height of the pile foundation, tidal range and mud surface line position, and different disease characteristic weights are assigned to different functional zones. S4: Identify corrosion, cracks, coating peeling and marine organism attachment in each functional zone to obtain image feature parameters of various diseases. Use the image feature parameters combined with the disease feature weights of each functional zone to output the disease severity index of each functional zone. S5: Register the current inspection image and the historical inspection image to the same pile foundation cylindrical coordinate system, calculate the change in the area ratio of various types of defects, and obtain the defect development trend value of each functional zone; S6: Convert the disease severity index and the disease development trend value into the overall risk index of the pile foundation, and combine them with the image credibility to fill the pre-edited report template to generate a pile foundation inspection report.

2. The method for identifying and detecting defects in offshore wind turbine pile foundations based on image processing according to claim 1, characterized in that: Image credibility Retrieval logic; ； in Indicates the image quality score. Indicates the scale resolution score. , These are the weighting coefficients, and their sum is one. Image quality score It is obtained by weighting the sharpness score, brightness score, contrast score, and occlusion score; The sharpness score is obtained by normalizing the Laplacian variance of the grayscale image of the inspection image; The brightness score is determined by the deviation of the average brightness of the target area from the preset target brightness median value; The contrast ratio is obtained by normalizing the standard deviation of the target area brightness relative to the preset minimum effective contrast ratio and the upper limit of the contrast ratio; The occlusion score is determined by the ratio of the pixel area of ​​the occluded region to the total area of ​​the target recognition region; The scale resolution score is obtained by normalizing the actual resolution of the target area relative to the preset high-precision recognition resolution and the upper limit of acceptable resolution. When the image confidence level is lower than the first confidence threshold, the output image is not suitable for direct labeling; when the image confidence level is between the first and second confidence thresholds, manual review of the labeling is triggered; when the image confidence level is higher than the second confidence threshold, the calculation of the deterministic disease level is initiated.

3. The method for identifying and detecting defects in offshore wind turbine pile foundations based on image processing according to claim 2, characterized in that: The pile foundation surface is divided into an atmospheric zone, a splash and tidal zone, a fully immersed zone, and a mud-surface adjacent zone. Obtain the axial height of the pile foundation Historical high tide level Historical low Mud surface height ; Mud surface adjacent area Defined as: ;in The width of the adjacent area of ​​the mud surface is half its width; splash zone Defined as: ;in This is a pre-set margin for tidal range extension; To extend the pre-set wave splash; Atmospheric region Defined as: ; Full immersion area Defined as: ; The disease weights corresponding to each functional zone can be set as follows: i = 1, 2, 3, or 4; where These respectively represent corrosion, cracks, coating peeling, and marine organism attachment in the zones. Weights within.

4. The method for identifying and detecting defects in offshore wind turbine pile foundations based on image processing according to claim 3, characterized in that: Within each functional zone, corrosion, cracks, coating peeling, and marine organism attachment are identified, and a disease severity index is output. Let the pixel set of the i-th functional partition be denoted as Let the identification mask for the m-th type of disease be denoted as ,in ; These respectively represent corrosion, cracks, coating peeling, and marine organism attachment; For pixel p, if it belongs to the m-th type of defect, then =1, if it does not belong to then =0; The disease area of ​​the m-th type of disease within the i-th functional zone is obtained through... The calculation yielded: in, represents the actual area of ​​the m-th type of disease within the i-th functional zone; ap represents the actual area of ​​the pile foundation surface corresponding to pixel p. Based on the actual affected area of ​​various diseases and the effective detection area of ​​the i-th functional zone, calculate the corrosion ratio, crack ratio, coating peeling ratio and marine organism attachment ratio. Multiply the area percentage of each diseased area by the corresponding disease weight in the i-th functional zone and sum them to obtain the disease severity index Si of the i-th functional zone.

5. The method for identifying and detecting defects in offshore wind turbine pile foundations based on image processing according to claim 4, characterized in that: The actual area of ​​the pile foundation surface is determined based on the cylindrical coordinates of the pile foundation after the pixel point is mapped; if the cylindrical coordinates of pixel point p are... The coordinate differences between its adjacent pixels in the circumferential and axial directions are respectively and ,but: ; Where D represents the pile diameter; D / 2 represents the pile radius; This represents the angular interval corresponding to pixel p in the circumferential direction; This represents the actual height interval corresponding to pixel p in the axial height direction. This represents the actual arc length of the pixel along the circumference of the pile foundation.

6. The method for identifying and detecting defects in offshore wind turbine pile foundations based on image processing according to claim 5, characterized in that: Register the current inspection image with the historical inspection image to the same pile foundation cylindrical coordinate system and calculate the disease development trend value; Calculate the disease development trend value of the i-th functional zone. : ； in, This represents the disease development trend value for the i-th functional zone; , , as well as These represent the development scores of various diseases within the i-th functional zone.

7. The method for identifying and detecting defects in offshore wind turbine pile foundations based on image processing according to claim 6, characterized in that: The logic for obtaining development scores; The current inspection time is denoted as Tc, the inspection time of the historical inspection image is denoted as Th, and the time interval between the two is calculated and denoted as Tg. Extract the corrosion percentage, crack percentage, coating peeling percentage, and marine organism attachment percentage of the target historical pile foundation corresponding to each functional zone from historical inspection images; Calculate the change in the area percentage of the m-th type of disease within the i-th functional zone: ; and These represent the area percentage of the m-th type of disease in the i-th functional zone at the current inspection time and the area percentage of the m-th type of disease in the i-th functional zone at the historical inspection time, respectively. Calculate the rate of disease spread per unit time: ; The changes in various diseases within the i-th functional zone are normalized to obtain the development score of the m-th disease. : ； in, This represents the development score of the m-th type of disease within the i-th functional zone; This represents the low development rate threshold for the m-th type of disease; This represents the high development rate threshold for the m-th type of disease; clip(·, 0, 1) restricts the calculation result to between 0 and 1.

8. The method for identifying and detecting defects in offshore wind turbine pile foundations based on image processing according to claim 7, characterized in that: The disease development trend value and disease severity index are converted into the overall risk index of the pile foundation; Let Si be the disease severity index of the i-th functional zone, combined with the disease development trend value of the i-th functional zone. Using the formula The partition risk value Ri of the i-th functional partition is calculated; where The weighting coefficients represent the severity index of the disease. β represents the weighting coefficient of the disease development trend value; Overall risk index of pile foundation .

9. An image processing-based offshore wind turbine pile foundation defect identification and detection system, applied to any one of the image processing-based offshore wind turbine pile foundation defect identification and detection methods described in claims 1-8, characterized in that: The image acquisition and coordinate mapping module is used to acquire inspection images, pile structure parameters and historical inspection parameters of offshore wind turbine pile foundations, and map the pixel positions in the inspection images to the cylindrical coordinate system of the pile foundation. The image credibility calculation module is used to calculate the image credibility based on the image quality score and scale resolution score of the inspection image, and output the image credibility as an unsuitable for direct judgment, manual review, or a deterministic disease level calculation instruction. The functional zoning and weight configuration module is used to divide the pile foundation surface into atmospheric zone, splash tidal zone, fully immersed zone and mud surface adjacent zone according to the pile foundation axial height, tidal range and mud surface line position, and to configure the disease weights of corrosion, cracking, coating peeling and marine organism attachment for different functional zones. The disease identification and severity calculation module is used to identify corrosion, cracks, coating peeling and marine organism attachment in each functional zone, generate corresponding disease masks, calculate the area ratio of each type of disease, and output the disease severity index of each functional zone in combination with the disease weight. The historical registration and trend analysis module is used to register the current inspection image and the historical inspection image to the same pile foundation cylindrical coordinate system, calculate the change in the area ratio of various diseases in the overlapping effective area, and output the disease development trend value of each functional zone. The risk calculation and report generation module is used to calculate the overall risk index of the pile foundation based on the severity index, trend value, structural importance coefficient and effective inspection area of ​​each functional zone, and generate a pile foundation inspection report in combination with image credibility.

Images