Dam-based defect detection method and system
By using feature recognition and representative point comparison technology, the problem of misjudgment in dam inspection has been solved, the detection accuracy of defects such as slope bulging has been improved, and the influence of external interference has been reduced.
Patent Information
- Application Number
- CN202511123796.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-12
- Publication Date
- 2025-11-21
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing drone aerial photography technology is easily affected by foreign objects in the inspection of dam slopes, resulting in insufficient detection accuracy and difficulty in effectively identifying defects such as slope bulging.
Feature recognition technology is used to identify suspected defect areas. By comparing the actual and theoretical positions of representative points, the point deviation parameters are calculated. Combined with the overall deviation parameters and invalid area analysis, it is determined whether the area is an actual defect area.
It improves the accuracy of dam defect detection, especially the identification of slope bulging, reduces misjudgments and external interference, and enhances the reliability of detection.
Smart Images
Figure CN120997183A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the field of water conservancy detection technology, in particular to a dam defect detection method and system. BACKGROUND
[0002] As important water conservancy infrastructure, the structural stability of the dam is directly related to flood control safety and water resources regulation. However, due to long-term influence of factors such as hydraulic erosion, settlement deformation, material aging, etc., the dam may have various defects, such as cracks, leakage, landslides, slope bulging, etc. Among them, slope bulging is a typical dam deformation phenomenon, which is manifested as local bulging of the dam body slope, usually caused by internal soil softening, increased seepage pressure or construction quality problems. If not discovered and treated in time, slope bulging may further evolve into landslides or even dam break accidents, threatening the safety of the downstream.
[0003] At present, the detection of dam slope bulging mainly relies on unmanned aerial vehicle aerial photography technology, combined with three-dimensional modeling and image analysis means for deformation monitoring. However, the unmanned aerial vehicle aerial photography is easily disturbed by slope foreign matters (such as temporary soil piles, construction materials, vegetation coverage, etc.), which may lead to misjudgment of non-structural bulging as slope bulging, affecting the detection accuracy, and there is still room for improvement. SUMMARY
[0004] In order to improve the accuracy of dam defect detection, the present application provides a dam defect detection method and system.
[0005] In the first aspect, the present application provides a dam defect detection method, which adopts the following technical scheme: A dam defect detection method, comprising: obtaining a dam surface image; performing feature recognition on the dam surface image to determine a suspected defect area, and determining a direct influence area in a preset dam division area according to the suspected defect area; determining the actual position of the preset representative point on each direct influence area in the dam surface image; comparing the actual position with a preset theoretical position of the representative point to determine a point deviation parameter, and calculating all point deviation parameters to determine an overall deviation parameter; judging whether the overall deviation parameter is greater than a preset reference deviation parameter; if the overall deviation parameter is greater than the reference deviation parameter, defining the suspected defect area as an actual defect area; if the overall deviation parameter is not greater than the reference deviation parameter, defining the suspected defect area as an external interference area.
[0006] Optionally, after the direct influence area is determined, the dam-based defect detection method further comprises: defining the direct influence area representing the point on the suspected defect area as an invalid reference area; counting according to the direct influence area to determine a direct influence quantity, and counting according to the invalid reference area to determine an invalid reference quantity; calculating according to the direct influence quantity and the invalid reference quantity to determine an invalid area proportion; judging whether the invalid area proportion is less than a preset full load proportion; if the invalid area proportion is not less than the full load proportion, outputting a signal for rechecking according to the current suspected defect area; if the invalid area proportion is less than the full load proportion, determining an evaluation true level corresponding to the invalid area proportion according to a preset evaluation matching relationship, and synchronously outputting the evaluation true level according to the definition of the suspected defect area.
[0007] Optionally, if the invalid area proportion is not less than the full load proportion, the dam-based defect detection method further comprises: determining an adjacent indirect influence area in the dam divided area according to the direct influence area, and defining the representative actual position of the representative point on the indirect influence area as an indirect actual position; calculating according to the indirect actual position and the corresponding representative theoretical position to determine a position deviation parameter; judging whether the position deviation parameter is greater than a preset required deviation parameter; if the position deviation parameter is not greater than the required deviation parameter, outputting a signal for rechecking according to the current suspected defect area; if the position deviation parameter is greater than the required deviation parameter, defining the suspected defect area as an actual defect area, and synchronously outputting the corresponding evaluation true level.
[0008] Optionally, the step of comparing the representative actual position with the preset representative theoretical position to determine a point position deviation parameter comprises: determining a contour separation distance according to the representative actual position and each contour point on the suspected defect area, defining the contour separation distance with the smallest value as an attractive separation distance, and constructing an attractive straight line with the contour point on the suspected defect area corresponding to the attractive separation distance and the representative actual position; projecting the representative actual position and the preset representative theoretical position on the attractive straight line to determine a position moving distance; determining a reference moving distance corresponding to the attractive separation distance according to a preset reference matching relationship; calculating according to the position moving distance and the reference moving distance to determine the point position deviation parameter.
[0009] Optionally, after the position moving distance is determined, the dam-based defect detection method further comprises: determining whether the direct influence area where the current representative point is located corresponds to at least two suspected defect areas; if the direct influence area where the current representative point is located does not correspond to at least two suspected defect areas, maintaining the determined position moving distance; if the direct influence area where the current representative point is located corresponds to at least two suspected defect areas, determining the straight line intersection angle according to the current attractive straight line and any one of the remaining attractive straight lines, and calculating the corresponding reference moving distance according to the straight line intersection angle to determine the external influence distance; correcting the current position moving distance according to the external influence distance.
[0010] Optionally, the method further comprises a step of determining a reference deviation parameter, which comprises: acquiring the detection pixel height of each pixel point in the suspected defect area; calculating the defect pixel height according to the preset plane pixel height and the detection pixel height, and calculating the defect actual height according to the preset conversion ratio and the defect pixel height; randomly selecting one of the defect actual heights as the center actual height, and defining the remaining defect actual heights as peripheral actual heights; calculating the center representative parameter according to the center actual height and the peripheral actual heights, and defining the center actual height corresponding to the center representative parameter with the largest value as the representative actual height; calculating the bulging height range according to the representative actual height and the preset proximity range, and counting the number of defect actual heights within the bulging height range to determine the degree effective number; determining the reference deviation parameter corresponding to the degree effective number and the representative actual height according to the preset degree matching relationship.
[0011] In a second aspect, the application provides a dam-based defect detection system, which adopts the following technical solution: A dam-based defect detection system comprises: an acquisition module for acquiring dam surface images; a processing module connected with the acquisition module and the judgment module, for storing and processing information; a judgment module connected with the acquisition module and the processing module, for judging information; the processing module performs feature recognition on the dam surface images to determine suspected defect areas, and determines direct influence areas in the preset dam division areas according to the suspected defect areas; The processing module determines the actual representative positions of the preset representative points on each direct influence area in the dam surface image; The processing module compares the actual representative positions with the preset theoretical representative positions to determine point deviation parameters, and calculates the overall deviation parameter based on all the point deviation parameters; The judging module judges whether the overall deviation parameter is greater than a preset reference deviation parameter; If the judging module judges that the overall deviation parameter is greater than the reference deviation parameter, the processing module defines the suspected defect area as an actual defect area; If the judging module judges that the overall deviation parameter is not greater than the reference deviation parameter, the processing module defines the suspected defect area as an external interference area.
[0012] In summary, the present application includes at least one of the following beneficial technical effects: When detecting the bulging condition of the dam slope, whether the bulging condition actually exists can be determined according to the movement of the points around the bulging area, thereby improving the accuracy of defect detection; The situation that multiple suspected defect areas are close to each other can be analyzed, further improving the accuracy of defect detection; The bulging degree of the bulging area is analyzed to effectively determine whether the bulging requirement is met, thereby further improving the accuracy of defect detection. BRIEF DESCRIPTION OF DRAWINGS
[0013] Figure 1 is a flowchart of a dam defect detection method.
[0014] Figure 2 is a schematic diagram of dam region division and suspected defect area position.
[0015] Figure 3 is a schematic diagram of the positions of multiple suspected defect areas interacting with each other.
[0016] Figure 4 is a module flowchart of a dam defect detection method. DETAILED DESCRIPTION
[0017] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application. Figures 1-4 The present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.
[0018] The present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.
[0019] The present application discloses a dam defect detection method, which refers to Figure 1, the method flow of the dam defect detection method comprises the following steps: Step S100: obtaining a dam surface image.
[0020] The dam surface image is a regional image of a position on the dam that needs to be detected for defects, obtained by a aerial unmanned aerial vehicle.
[0021] Step S101: performing feature recognition on the dam surface image to determine a suspected defect region, and determining a direct impact region in a preset dam division region according to the suspected defect region.
[0022] The suspected defect region is a region with a protrusion obtained after three-dimensional modeling analysis and image analysis of the obtained dam surface image. The specific feature recognition method is set by the staff according to the actual situation. The feature recognition method is a routine technical means for those skilled in the art, which is not described here. The dam division region is a fixed region previously divided on the dam surface, wherein each region is arranged in an array, and each region is a square. The specific setting is made by the staff according to the actual situation. The direct impact region is the dam division region that has an intersection with the suspected defect region, which can be understood with reference to Figure 2 .
[0023] Step S102: determining the actual position of a preset representative point on each direct impact region in the dam surface image.
[0024] The representative point is a position point on each direct impact region that can be identified by image recognition, for example, a point position that is different from the rest of the positions in color. The specific setting can be made by the staff. The actual position is the position of the representative point on the actual two-dimensional surface, i.e. the dam surface, determined according to the position of the unmanned aerial vehicle when obtaining the dam surface image.
[0025] Step S103: comparing the actual position and the preset theoretical position to determine a point position deviation parameter, and calculating the overall deviation parameter according to all point position deviation parameters.
[0026] The theoretical position is the position coordinate point that the representative point needs to be in when the dam surface has not changed. The point position deviation parameter is a parameter value reflecting the deviation of the representative point. The value can be the distance between the actual position and the theoretical position, or it can be determined with reference to steps S400-S403. The overall deviation parameter is the average value of all point position deviation parameters.
[0027] Step S104: determining whether the overall deviation parameter is greater than a preset reference deviation parameter.
[0028] The reference deviation parameter is a minimum overall deviation parameter required to be met when the staff sets the deviation of the direct influence area corresponding to the suspected defect area. The purpose of the judgment is to know whether the surrounding area of the current suspected defect area has a deviation, so as to determine whether the current protruding area is a slope bulge.
[0029] Step S1041: If the overall deviation parameter is greater than the reference deviation parameter, the suspected defect area is defined as an actual defect area.
[0030] When the overall deviation parameter is greater than the reference deviation parameter, it means that there is a slope bulge at present, so it can be defined as an actual defect area.
[0031] Step S1042: If the overall deviation parameter is not greater than the reference deviation parameter, the suspected defect area is defined as an external interference area.
[0032] When the overall deviation parameter is not greater than the reference deviation parameter, it means that the surrounding area of the current suspected defect area has no obvious deviation change, so the area basically does not have a slope bulge, so it can be defined as an external interference area, thereby improving the accuracy of the dam defect detection.
[0033] After the direct influence area is determined, the dam defect detection method further includes: Step S200: Defining the direct influence area where the representative point is on the suspected defect area as an invalid reference area.
[0034] The invalid reference area is defined to identify the direct influence area that cannot be used for reference, for subsequent analysis.
[0035] Step S201: Counting the direct influence areas according to the direct influence areas to determine the direct influence number, and counting the invalid reference areas according to the invalid reference areas to determine the invalid reference number.
[0036] The direct influence number is the total number of the determined direct influence areas, and the invalid reference number is the total number of the determined invalid reference areas.
[0037] Step S202: Calculating the invalid area ratio according to the direct influence number and the invalid reference number.
[0038] The invalid area ratio is the ratio of the number of all invalid reference areas to all direct influence areas, which is determined by dividing the invalid reference number by the direct influence number.
[0039] Step S203: Judging whether the invalid area ratio is less than a preset full load ratio.
[0040] The full load ratio is the minimum invalid area ratio required for the worker to determine the suspected defect area by directly affecting the area. The full load ratio is generally 100%. The purpose of the determination is to determine whether the representative point on the directly affected area can be used to determine the protruding abnormality.
[0041] Step S2031: If the invalid area ratio is not less than the full load ratio, output a waiting review signal according to the current suspected defect area.
[0042] When the invalid area ratio is not less than the full load ratio, there is no directly affected area available for analysis. Therefore, a waiting review signal is output for subsequent manual review by the worker.
[0043] Step S2032: If the invalid area ratio is less than the full load ratio, determine the evaluation true level corresponding to the invalid area ratio according to the preset evaluation matching relationship, and output the evaluation true level synchronously according to the definition of the suspected defect area.
[0044] When the invalid area ratio is less than the full load ratio, there is a directly affected area available for protruding abnormality analysis. Therefore, the suspected defect area analysis is performed normally. The evaluation true level is a level value reflecting the reliability of the determination. The higher the level, the stronger the reliability. Different invalid area ratios correspond to different evaluation true levels. The evaluation matching relationship between them is determined by the worker in advance and recorded in storage. It is necessary to ensure that the lower the invalid area ratio, the higher the corresponding evaluation true level.
[0045] If the invalid area ratio is not less than the full load ratio, the dam defect detection method further comprises: Step S300: Determine the adjacent indirect affected area in the dam division area according to the directly affected area, and define the representative actual position of the representative point on the indirect affected area as the indirect actual position.
[0046] The indirect affected area is the dam division area adjacent to any one of the directly affected areas in the dam division area. The indirect actual position is the position point of the representative point on the indirect affected area.
[0047] Step S301: Calculate the position deviation parameter according to the indirect actual position and the corresponding representative theoretical position.
[0048] The position deviation parameter is a parameter value reflecting the deviation of the representative point on the indirect affected area. The specific determination method is consistent with the overall deviation parameter, which is not described here.
[0049] Step S302: Determine whether the position deviation parameter is greater than the preset required deviation parameter.
[0050] The requirement deviation parameter is a minimum position deviation parameter required to be reached by the staff when determining the bulging condition of the suspected defect area through the convexity of the suspected defect area directly affecting the indirect influence area. The purpose of the determination is to know whether the bulging condition of the suspected defect area can be directly determined.
[0051] Step S3021: If the position deviation parameter is not greater than the requirement deviation parameter, output a waiting review signal according to the current suspected defect area.
[0052] When the position deviation parameter is not greater than the requirement deviation parameter, it means that the bulging condition cannot be directly determined, and at this time, the waiting review signal is output for manual review.
[0053] Step S3022: If the position deviation parameter is greater than the requirement deviation parameter, define the suspected defect area as an actual defect area, and output the corresponding evaluation true level synchronously.
[0054] When the position deviation parameter is greater than the requirement deviation parameter, it means that the suspected defect area has a bulging condition in the actual situation, and therefore it can be determined as an actual defect area, thereby reducing the amount of subsequent review required.
[0055] The step of comparing the representative actual position with the preset representative theoretical position to determine the point position deviation parameter includes: Step S400: Determine the contour separation distance according to the representative actual position and each contour point on the suspected defect area, define the contour separation distance with the smallest value as the attractive separation distance, and construct an attractive straight line with the contour point on the suspected defect area corresponding to the attractive separation distance and the representative actual position.
[0056] The contour separation distance is the straight line distance value of the representative actual position and the contour point on the suspected defect area in the two-dimensional plane. The attractive separation distance is defined to identify the contour separation distance that is most likely to pull the representative point, facilitating subsequent analysis. The attractive straight line is constructed to facilitate determination of the pulling direction of the suspected defect area on the representative point, facilitating subsequent analysis.
[0057] Step S401: Project the representative actual position and the preset representative theoretical position on the attractive straight line to determine the position movement distance.
[0058] The position movement distance is the straight line distance value between the projection points obtained by projecting the representative actual position and the representative theoretical position on the attractive straight line.
[0059] Step S402: Determine the reference movement distance corresponding to the attractive separation distance according to the preset reference matching relationship.
[0060] The reference moving distance is a representative point deviation distance required to be reached under a normal condition at a current attraction separation distance. Different attraction separation distances result in different pulling forces, and at this time, the corresponding reference moving distances are also different. A reference matching relationship between the two is set by a worker according to actual conditions, and details are not described herein.
[0061] Step S403: calculating to determine a point position deviation parameter according to the position moving distance and the reference moving distance.
[0062] The point position deviation parameter is a parameter value obtained by dividing the position moving distance by the reference moving distance.
[0063] After the position moving distance is determined, the dam defect detection method further includes: Step S500: judging whether the current representative point is located in a direct influence area corresponding to at least two suspected defect areas.
[0064] The purpose of the judgment is to know whether a single representative point is pulled by two convex positions.
[0065] Step S5001: if the current representative point is located in a direct influence area not corresponding to at least two suspected defect areas, maintaining the determined position moving distance.
[0066] When the current representative point is located in a direct influence area not corresponding to at least two suspected defect areas, it is indicated that the current representative point is pulled by only one suspected defect area, and thus normal analysis can be performed according to the determined position moving distance.
[0067] Step S5002: if the current representative point is located in a direct influence area corresponding to at least two suspected defect areas, determining a straight line intersection angle according to a current attraction straight line and any one of the remaining attraction straight lines, and calculating a reference moving distance corresponding to the straight line intersection angle to determine an external influence distance.
[0068] When the current representative point is located in a direct influence area corresponding to at least two suspected defect areas, it is indicated that the current representative point is pulled by at least two suspected defect areas, and thus further analysis is required. The straight line intersection angle is an angle formed by two attraction straight lines, and the external influence distance is a projection distance value of a reference moving distance of a suspected defect area corresponding to another attraction straight line on the current attraction straight line. The value reflects the pulling influence of the external remaining suspected defect area on the current representative point. When the angle of the straight line intersection angle is greater than 90°, the corresponding external influence distance is a negative value, and vice versa. Figure 3
[0069] Step S501: correcting the current position moving distance according to the external influence distance.
[0070] The current position moving distance is corrected by adding the external influence distance to the current position moving distance, so as to improve the accuracy of the current position moving distance, and further improve the accuracy of the defect condition analysis.
[0071] The method further comprises a step of determining the reference deviation parameter, and the step comprises: Step S600: Obtain the detection pixel height of each pixel point in the suspected defect region.
[0072] The detection pixel height is the pixel height of each pixel point in the suspected defect region in the image, which can be obtained by the binocular camera on the dam surface image.
[0073] Step S601: Calculate the defect pixel height according to the preset plane pixel height and the detection pixel height, and calculate the defect actual height according to the preset conversion ratio and the defect pixel height.
[0074] The plane pixel height is the height value that each pixel point needs to be in when the dam surface has no foreign matter, and the defect pixel height is determined by subtracting the plane pixel height from the detection pixel height; the conversion ratio is the height conversion parameter in the image and the actual situation, and the actual height value of each position point on the defect detection region can be obtained by multiplying the defect pixel height by the conversion ratio, that is, the defect actual height.
[0075] Step S602: Randomly select one of the defect actual heights as the center actual height, and define the remaining defect actual heights as the peripheral actual heights.
[0076] The center actual height and the peripheral actual height are defined to distinguish different defect actual heights for subsequent analysis.
[0077] Step S603: Calculate the center representative parameter according to the center actual height and each peripheral actual height, and define the center actual height corresponding to the center representative parameter with the largest value as the representative actual height.
[0078] The center representative parameter is a reliable parameter value that the center actual height can reflect the current convexity, and the larger the value is, the more the corresponding center actual height can represent the convexity, wherein the calculation method of the center representative parameter is as follows: subtract each center actual height from each peripheral actual height and add the absolute values, and then calculate the reciprocal of the sum to determine; the representative actual height is defined to identify the parameter value of the current convexity, which is convenient for subsequent analysis.
[0079] Step S604: according to the representative actual height and the preset proximity range to calculate to build the bulging height range, and according to the actual height of the defect in the bulging height range to determine the degree effective quantity.
[0080] The proximity range is a range of the difference allowed to appear between the parameter set by the staff and the representative actual height, and the representative actual height, and the bulging height range can be determined by adding the upper end point and the lower end point of the proximity range to the representative actual height; The degree effective quantity is the number value of the actual height of the defect in the bulging height range, that is, the number of position points that reach the corresponding height.
[0081] Step S605: according to the preset degree matching relationship to determine the degree effective quantity and the reference deviation parameter corresponding to the representative actual height.
[0082] Different degree effective quantities indicate different overall protrusion quantities, different representative actual heights indicate different overall protrusion conditions, and therefore the pulling ability of each representative point of the surrounding area is different, so that the corresponding reference deviation parameter is different, and the degree matching relationship between the three is determined by the staff through multiple tests in advance, which needs to be guaranteed. The greater the degree effective quantity and the greater the representative actual height, the greater the corresponding reference deviation parameter.
[0083] Reference Figure 4 Based on the same inventive concept, the embodiment of the present application provides a dam-based defect detection system, comprising: An acquisition module is configured to acquire a dam surface image. A processing module is connected with the acquisition module and the judgment module, and is configured to store and process information. A judgment module is connected with the acquisition module and the processing module, and is configured to judge information. The processing module performs feature recognition on the dam surface image to determine a suspected defect area, and determines a direct impact area in a preset dam division area according to the suspected defect area. The processing module determines the representative actual position of each representative point in the direct impact area in the dam surface image. The processing module compares the representative actual position with a preset representative theoretical position to determine a point position deviation parameter, and calculates an overall deviation parameter according to all point position deviation parameters. The judgment module judges whether the overall deviation parameter is greater than a preset reference deviation parameter. If the judgment module judges that the overall deviation parameter is greater than the reference deviation parameter, the processing module defines the suspected defect area as an actual defect area. If the judgment module judges that the overall deviation parameter is not greater than the reference deviation parameter, the processing module defines the suspected defect area as an external interference area. a direct influence area analysis module for analyzing the determined direct influence area; a review situation analysis module for analyzing whether a review is needed for a case without a suitable direct influence area; a point deviation parameter calculation module for calculating a relatively accurate point deviation parameter; a position movement distance correction module for correcting a position movement distance through multiple defect area analysis; a reference deviation parameter determination module for determining a relatively suitable reference deviation parameter for data analysis.
[0084] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional modules is taken as an example for illustration, and in actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiments, which will not be described here.
Claims
1. A defect detection method based on dams, characterized in that, include: Acquire images of the dam surface; Feature recognition is performed on the dam surface image to identify suspected defect areas, and the directly affected area is determined based on the suspected defect areas within the preset dam division area; Determine the actual locations of pre-defined representative points on each directly affected area from the dam surface image; The point deviation parameters are determined by comparing the actual position with the preset theoretical position, and the overall deviation parameters are determined by calculating based on all the point deviation parameters. Determine whether the overall deviation parameter is greater than the preset benchmark deviation parameter; If the overall deviation parameter is greater than the baseline deviation parameter, then the suspected defect area is defined as the actual defect area. If the overall deviation parameter is not greater than the baseline deviation parameter, the suspected defect area is defined as the external interference area.
2. The defect detection method based on dams according to claim 1, characterized in that, After the directly affected area is determined, the defect detection methods based on dams also include: The area directly affected by the representative point located in the suspected defect area is defined as the invalid reference area. The number of direct impacts is determined by counting the areas directly affected, and the number of invalid references is determined by counting the areas of invalid references. The percentage of invalid areas is determined by calculating the number of directly affected areas and the number of invalid reference areas. Determine if the percentage of invalid areas is less than the preset percentage of full load; If the percentage of invalid areas is not less than the percentage of full load, then output a wait-for-review signal based on the current suspected defective areas; If the proportion of invalid areas is less than the proportion of full load, the true evaluation level corresponding to the proportion of invalid areas is determined according to the preset evaluation matching relationship, and the true evaluation level is output synchronously according to the definition of suspected defective areas.
3. The defect detection method based on dams according to claim 2, characterized in that, If the proportion of invalid areas is not less than the proportion of full load, defect detection methods based on dams also include: Based on the direct impact area and the dike division area, the adjacent indirect impact area is determined, and the representative actual position of the representative point on the indirect impact area is defined as the indirect actual position; The position deviation parameter is determined by calculation based on the indirect actual position and the corresponding representative theoretical position. Determine whether the position deviation parameter is greater than the preset required deviation parameter; If the position deviation parameter is not greater than the required deviation parameter, then output a waiting-for-review signal based on the current suspected defect area; If the location deviation parameter is greater than the demand deviation parameter, the suspected defect area is defined as the actual defect area, and the corresponding evaluation true level is output synchronously.
4. The defect detection method based on dams according to claim 1, characterized in that, The steps for determining the point deviation parameter by comparing the actual representative location with the preset representative theoretical location include: The contour distance is determined based on the contour points representing the actual location and the suspected defect area. The contour distance with the smallest value is defined as the attraction distance. An attraction line is constructed by the contour points on the suspected defect area corresponding to the attraction distance and the actual location. The distance the position moves is determined by projecting the actual position and the pre-set theoretical position onto the attraction line. The baseline movement distance corresponding to the attraction interval is determined based on the preset baseline matching relationship; The point deviation parameter is determined by calculating the distance moved to the location and the reference distance.
5. The defect detection method based on dams according to claim 4, characterized in that, Once the location movement distance is determined, the defect detection method based on the dam also includes: Determine whether the directly affected area where the current representative point is located corresponds to at least two suspected defect areas; If the directly affected area where the current representative point is located does not correspond to at least two suspected defect areas, then the currently determined position movement distance will be maintained. If the area directly affected by the current representative point corresponds to at least two suspected defect areas, then the angle between the intersecting lines is determined based on the current attraction line and any other attraction line, and the corresponding reference movement distance is calculated based on the angle between the intersecting lines to determine the external influence distance. The distance the current position moves is corrected based on the distance of external influences.
6. The defect detection method based on dams according to claim 1, characterized in that, It also includes a step for determining the benchmark deviation parameter, which includes: Obtain the detection pixel height of each pixel in the suspected defect area; The defect pixel height is determined by calculation based on the preset planar pixel height and the detection pixel height, and the actual defect height is determined by calculation based on the preset conversion ratio and the defect pixel height. Randomly select one actual height of each defect as the center actual height, and define the actual heights of the remaining defects as the surrounding actual heights; The center representative parameter is determined by calculation based on the actual height of the center and the actual height of each surrounding area, and the actual height of the center corresponding to the center representative parameter with the largest value is defined as the representative actual height. The bulging height range is constructed by calculating based on the actual height and a preset similar range, and the effective number of defects within the bulging height range is determined by counting the actual height of the defects. The effective number of degrees and the benchmark deviation parameter corresponding to the actual height are determined based on the preset degree matching relationship.
7. A defect detection system based on dams, characterized in that, include: The acquisition module is used to acquire images of the dam surface; The processing module, connected to the acquisition and judgment modules, is used for information storage and processing; The judgment module, connected to the acquisition and processing modules, is used for judging information. The processing module performs feature recognition on the dam surface image to determine suspected defect areas, and determines the directly affected area within the preset dam division area based on the suspected defect areas; The processing module determines the actual location of the preset representative points on each directly affected area in the dam surface image; The processing module compares the actual position with the preset theoretical position to determine the position deviation parameter, and calculates the overall deviation parameter based on all position deviation parameters. The judgment module determines whether the overall deviation parameter is greater than the preset benchmark deviation parameter; If the judgment module determines that the overall deviation parameter is greater than the benchmark deviation parameter, the processing module defines the suspected defect area as the actual defect area. If the judgment module determines that the overall deviation parameter is not greater than the baseline deviation parameter, the processing module defines the suspected defect area as an external interference area.