A Method for Identifying Slope Landslide and Settlement Defects in Riverbank Protection Based on Image Composite Analysis
By using image composite analysis, we can identify slippage and settlement defects in riverbank protection, form the range of turf density and texture, and distinguish slippage and settlement areas. This solves the problem of low identification efficiency of riverbank protection and achieves efficient and accurate identification and repair.
Patent Information
- Application Number
- CN202611304667.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-26
- Publication Date
- 2026-11-10
AI Technical Summary
Existing technologies have low efficiency in identifying slippage and settlement defects in riverbank protection, and manual identification is time-consuming and labor-intensive, making it difficult to carry out efficient repairs.
An image-based composite analysis method is adopted to acquire riverbank protection images, eliminate interference, identify turf density ranges, form low-density and high-density blocks, aggregate textures, match slippage and settlement areas, and form settlement and slippage areas.
It improves the efficiency of identifying slippage and settlement defects in riverbank protection, ensures accurate identification, enables targeted repairs, and reduces the time required for manual intervention.
Smart Images

Figure CN123071301A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of water conservancy engineering technology, specifically to a method for identifying riverbank slope slippage and settlement defects based on image composite analysis. Background Technology
[0002] Riverbank ecological revetment is a water conservancy project that combines flood control and ecological restoration, aiming to maintain slope stability through vegetation or engineering materials. Typically, turf or stone mounds are used to stabilize the slopes on both sides of the river. Here, we only consider riverbank revetment formed by turf. Due to long-term erosion by rainwater, soil is prone to displacement or settlement, requiring restoration of the affected areas. However, rivers are usually very long, and manual identification is time-consuming and labor-intensive, with efficiency needing improvement. Summary of the Invention
[0003] To address the aforementioned technical problems, this paper provides a method for identifying riverbank slope slippage and settlement defects based on image composite analysis. This technical solution solves the problems mentioned in the background section.
[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for identifying riverbank landslide and settlement defects based on image composite analysis includes: Obtain images of riverbank revetments covered with turf, remove interference from the revetment images, and obtain surface-processed images. At least one normal region is identified in the surface-processed image, and the turf density range is analyzed within the normal region. Based on the turf density range, at least one low-density block and one high-density block in the surface-processed image are analyzed and obtained. The low-density block and the high-density block are aggregated respectively to form at least one low-density texture and one high-density texture. Low-density textures and high-density textures are respectively completed and polymerized to form low-polymer textures and high-polymer textures; High-polymer textures were matched with low-polymer textures, and low-polymer textures were classified into slip textures and settling textures. Based on the settlement patterns and slip patterns, at least one settlement area and at least one slip area are formed.
[0005] Preferably, obtaining images of riverbank revetments covered with turf includes the following steps: Select at least one marker point on the riverbank revetment and use the plane perpendicular to the riverbank revetment that passes through the marker point as the reference plane; At least one reference plane divides the riverbank protection into at least one local protection, and the local protection with the largest area is designated as the characteristic local protection. Take a shot perpendicular to the center of the feature slope protection, increase the distance between the camera and the feature slope protection until the feature slope protection is completely contained within the camera's field of view, and use the current distance between the camera and the feature slope protection as the preset distance; Take a picture perpendicular to the center of the local slope protection at a preset distance to obtain an overall image of the local slope protection. Delete the parts other than the local slope protection from the overall image of the local slope protection to obtain a reference image of the local slope protection. The reference images of the partial slope protection are stitched together sequentially to obtain the slope protection image of the river channel.
[0006] Preferably, the process of eliminating interference in the slope protection image to obtain a surface-processed image includes the following steps: The pixel values of the slope protection soil in the image are obtained in advance as the first pixel value, and the pixel values of the slope protection turf in the image are obtained in advance as the second pixel value; Pixels in the slope protection image whose pixel value is not equal to the first pixel value or the second pixel value are taken as points to be replaced; Adjacent points to be replaced are aggregated to obtain at least one area to be replaced. Feature lines perpendicular to the bottom of the slope protection image are generated on both sides of the area to be replaced. The feature line is shifted to a position tangent to the region to be replaced, and the distance between the two feature lines is taken as the feature distance. The area to be replaced is shifted to the left by a feature distance to obtain a reference area. The image covered by the reference area is used to replace the image in the area to be replaced. After the slope protection image is replaced, the surface-processed image is obtained.
[0007] Preferably, identifying at least one normal region in the surface-treated image includes the following steps: Based on historical data of riverbank protection, the locations where slippage or settlement has occurred in the riverbank protection are identified as abnormal areas. The parts of the surface-processed image other than the abnormal areas are identified as normal parts. The normal parts are then uniformly segmented to obtain at least one normal area.
[0008] Preferably, the step of analyzing the turf density range in a normal area includes the following steps: In the normal area, count the number of pixels whose pixel value is equal to the first pixel value, and obtain the first number; In the normal area, count the number of pixels whose pixel value is equal to the second pixel value to obtain the second number; The first number is superimposed with the second number to obtain the third number. The second number is divided by the third number to obtain the characteristic value of the normal area. The maximum and minimum values of the characteristic value of the normal area are used as endpoints to form the turf density range.
[0009] Preferably, the analysis to obtain at least one low-density block and one high-density block in the surface-processed image includes the following steps: The surface-processed image is uniformly segmented to obtain at least one distribution block. The number of pixels in the distribution block whose pixel value is equal to the second pixel value is counted and used as the fourth number. Divide the fourth number by the total number of pixels in the distribution block to obtain the distribution value of the distribution block; If the distribution value of a distribution area is less than the value at the left end of the turf density range, the distribution area is a low-density area; if the distribution value of a distribution area is greater than the value at the right end of the turf density range, the distribution area is a high-density area.
[0010] Preferably, the step of aggregating low-density blocks and high-density blocks to form at least one low-density texture and at least one high-density texture includes the following steps: If two low-density blocks are in contact, then the two low-density blocks are merged; if two high-density blocks are in contact, then the two high-density blocks are merged. The texture formed by merging low-density blocks is designated as low-density texture, and the texture formed by merging high-density blocks is designated as high-density texture.
[0011] Preferably, the formation of low-polymerization textures and the formation of high-polymerization textures include the following steps: The two ends of the low-density texture are designated as the low end, and the two ends of the high-density texture are designated as the high end. Select one of the low ends as the target low end, select the low end that is closest to the target low end as the feature low end, and select the high end that is closest to the target low end as the feature high end. If the distance from the bottom of the target to the bottom of the feature is less than the distance from the bottom of the target to the top of the feature, then connect the aggregated line segment between the bottom of the target and the bottom of the feature; otherwise, do not perform any processing. When the target low end traverses all low ends, the low-density texture and the texture formed by the connected low-density texture and the aggregated line segments are regarded as low-aggregate texture. Select one of the high-ends as the target high-end, the high-end closest to the target high-end as the benchmark high-end, and the low-end closest to the target high-end as the benchmark low-end; If the distance from the target high end to the base high end is less than the distance from the target high end to the base low end, then connect the convergent line segment between the target high end and the base high end; otherwise, do not perform any processing. When the target high-end traverses all the high-ends, the high-density texture and the texture formed by the aggregated line segments connected into one are called high-aggregate texture.
[0012] Preferably, the step of matching high-polymerization patterns with low-polymerization patterns and classifying low-polymerization patterns into slippage patterns and settling patterns includes the following steps: Match the high-polymer texture to the low-polymer texture that contacts both ends of it. If there is no low-polymer texture that contacts both ends of the high-polymer texture, match the low-polymer texture with the smallest sum of distances to both ends of the high-polymer texture to the high-polymer texture. Low-polymer patterns that match high-polymer patterns are designated as slip patterns, while low-polymer patterns that do not match high-polymer patterns are designated as settling patterns.
[0013] Preferably, forming at least one settlement area and at least one slippage area includes the following steps: The area enclosed by the settlement lines is defined as the settlement area; If a high-polymerization texture and its matching slip texture form a closed loop, then the area enclosed by the high-polymerization texture and its matching slip texture is taken as the local area, and the area covered by the local area and the slip texture matched by the high-polymerization texture is taken as the slip area. If the high-polymer texture and its matching slip texture do not form a closed loop, then connect the left end of the high-polymer texture and its matching slip texture to obtain the first line segment, and connect the right end of the high-polymer texture and its matching slip texture to obtain the second line segment. The area enclosed by the first line segment, the second line segment, the high-polymer texture, and the slip texture matched by the high-polymer texture is defined as the local area, and the area covered by the local area and the slip texture matched by the high-polymer texture is defined as the slip area.
[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: By eliminating interference in the slope protection image, obtaining the turf density range, forming low-polymer textures, forming high-polymer textures, and identifying settlement and slippage areas, aerial photography can be used to acquire images of riverbank slopes. By eliminating obstacle interference, an interference-free slope protection image is obtained, ensuring the accuracy of identification. During identification, the actual characteristics of slippage and settlement are distinguished, thereby identifying the location of slippage and settlement separately, and carrying out targeted repairs. The gaps in the textures that caused slippage and settlement are filled, thus accurately identifying the range of slippage and settlement. Because it is based on image visual recognition, the identification efficiency can be greatly improved. Attached Figure Description
[0015] Figure 1 This is a flowchart illustrating the method for identifying riverbank landslide and settlement defects based on image composite analysis according to the present invention. Figure 2 This is a schematic diagram of the process for obtaining images of a riverbank slope covered with turf according to the present invention. Figure 3 This is a schematic diagram illustrating the process of eliminating interference from slope protection images to obtain surface-processed images according to the present invention. Figure 4 This is a schematic diagram of the process for analyzing and obtaining the range of turf density in a normal area according to the present invention; Figure 5 This is a schematic diagram illustrating the process of obtaining at least one low-density block and a high-density block in a surface-processed image according to the present invention. Figure 6 This is a schematic diagram of the process of aggregating low-density blocks and high-density blocks to form at least one low-density texture and one high-density texture, respectively, according to the present invention. Figure 7 This is a schematic diagram illustrating the process of forming low-polymer textures and high-polymer textures according to the present invention. Figure 8 This is a schematic diagram illustrating the process of matching high-polymerization patterns with low-polymerization patterns and classifying low-polymerization patterns into slip patterns and settling patterns according to the present invention. Figure 9 This is a schematic diagram of the process for forming at least one settling region and at least one slip region according to the present invention. Detailed Implementation
[0016] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.
[0017] Reference Figure 1 As shown, the method for identifying riverbank landslide settlement defects based on image composite analysis includes: Obtain images of riverbank revetments covered with turf, remove interference from the revetment images, and obtain surface-processed images. At least one normal region is identified in the surface-processed image, and the turf density range is analyzed within the normal region. Based on the turf density range, at least one low-density block and one high-density block in the surface-processed image are analyzed and obtained. The low-density block and the high-density block are aggregated respectively to form at least one low-density texture and one high-density texture. Low-density textures and high-density textures are respectively completed and polymerized to form low-polymer textures and high-polymer textures; High-polymer textures were matched with low-polymer textures, and low-polymer textures were classified into slip textures and settling textures. Based on the settlement patterns and slip patterns, at least one settlement area and at least one slip area are formed.
[0018] Using aerial photography allows for faster movement than manual identification, and machine vision identification requires almost no additional time, thus improving efficiency. However, it's crucial to differentiate between landslides and settlement during identification, as their repair methods differ. Furthermore, riverbank revetments may contain debris such as rocks and leaves, which can interfere with the process. Therefore, these obstructions must be removed during identification, requiring a series of subsequent steps to address them.
[0019] Reference Figure 2 As shown, obtaining images of a riverbank revetment covered with turf includes the following steps: Select at least one marker point on the riverbank revetment and use the plane perpendicular to the riverbank revetment that passes through the marker point as the reference plane; At least one reference plane divides the riverbank protection into at least one local protection, and the local protection with the largest area is designated as the characteristic local protection. Take a shot perpendicular to the center of the feature slope protection, increase the distance between the camera and the feature slope protection until the feature slope protection is completely contained within the camera's field of view, and use the current distance between the camera and the feature slope protection as the preset distance; Take a picture perpendicular to the center of the local slope protection at a preset distance to obtain an overall image of the local slope protection. Delete the parts other than the local slope protection from the overall image of the local slope protection to obtain a reference image of the local slope protection. The reference images of the partial slope protection are stitched together sequentially to obtain the slope protection image of the river channel.
[0020] Here, due to the long length of the riverbank revetment and the limited field of view of the camera, only partial images can be captured each time. In order to ensure the consistency of the size in the images, the shooting distance needs to be consistent. At the same time, in order to facilitate the stitching of the captured images, it is necessary to set marker points in the riverbank revetment and use the marker points as references to stitch the images together, thereby obtaining the revetment image of the riverbank revetment.
[0021] Reference Figure 3 As shown, the process of removing interference from the slope protection image to obtain a surface-processed image includes the following steps: The pixel values of the slope protection soil in the image are obtained in advance as the first pixel value, and the pixel values of the slope protection turf in the image are obtained in advance as the second pixel value; Pixels in the slope protection image whose pixel value is not equal to the first pixel value or the second pixel value are taken as points to be replaced; Adjacent points to be replaced are aggregated to obtain at least one area to be replaced. Feature lines perpendicular to the bottom of the slope protection image are generated on both sides of the area to be replaced. The feature line is shifted to a position tangent to the region to be replaced, and the distance between the two feature lines is taken as the feature distance. The area to be replaced is shifted to the left by a feature distance to obtain a reference area. The image covered by the reference area is used to replace the image in the area to be replaced. After the slope protection image is replaced, the surface-processed image is obtained.
[0022] Since the riverbank protection in this scheme uses turf, in the absence of debris, the turf alone can interfere with the identification process. Therefore, the turf near the debris is used to replace it. The replacement turf is similar to the turf actually covered by the debris, so it will not affect the subsequent identification results. Turf is composed of soil and grass, and the gaps between the grass are filled with soil. Therefore, the pixel values of soil and grass are known. Thus, pixels that do not have these two pixel values are treated as debris pixels. In this way, the location of the debris can be identified and replaced, thereby eliminating interference.
[0023] Identifying at least one normal region in a surface-processed image includes the following steps: Based on historical data of riverbank protection, the locations where slippage or settlement has occurred in the riverbank protection are identified as abnormal areas. The parts of the surface-processed image other than the abnormal areas are identified as normal parts. The normal parts are then uniformly segmented to obtain at least one normal area.
[0024] Reference Figure 4 As shown, the analysis of the turf density range in the normal area includes the following steps: In the normal area, count the number of pixels whose pixel value is equal to the first pixel value, and obtain the first number; In the normal area, count the number of pixels whose pixel value is equal to the second pixel value to obtain the second number; The first number is superimposed with the second number to obtain the third number. The second number is divided by the third number to obtain the characteristic value of the normal area. The maximum and minimum values of the characteristic value of the normal area are used as endpoints to form the turf density range.
[0025] When settlement or slippage occurs, the density of turf at the edge of the area changes. During settlement, the outline of the settlement area sinks and is stretched, increasing the spacing between the grass blades. As a result, more soil is exposed, making the turf density lower than in the normal area. During slippage, assuming turf A slides down due to gravity, the turf under turf A accumulates, resulting in a higher density than in the normal area. Due to the downward movement of turf A, soil is exposed on the upper side of turf A, creating a patch where the grass density is lower than in the normal area. On both sides of turf A, soil is brought out along with the movement, causing the grass density on both sides of turf A to be lower than in the normal area. Therefore, it is necessary to obtain at least one normal area that has not experienced slippage or subsidence based on historical data, form a turf density range by the turf density in the normal area, and use the turf density range to identify subsidence or slippage.
[0026] Reference Figure 5 As shown, the analysis to obtain at least one low-density block and one high-density block in the surface-processed image includes the following steps: The surface-processed image is uniformly segmented to obtain at least one distribution block. The number of pixels in the distribution block whose pixel value is equal to the second pixel value is counted and used as the fourth number. Divide the fourth number by the total number of pixels in the distribution block to obtain the distribution value of the distribution block; If the distribution value of a distribution area is less than the value at the left end of the turf density range, the distribution area is a low-density area; if the distribution value of a distribution area is greater than the value at the right end of the turf density range, the distribution area is a high-density area.
[0027] Reference Figure 6 As shown, the process of aggregating low-density blocks and high-density blocks to form at least one low-density texture and one high-density texture includes the following steps: If two low-density blocks are in contact, then the two low-density blocks are merged; if two high-density blocks are in contact, then the two high-density blocks are merged. The texture formed by merging low-density blocks is designated as low-density texture, and the texture formed by merging high-density blocks is designated as high-density texture.
[0028] Reference Figure 7 As shown, forming low-polymer textures and forming high-polymer textures include the following steps: The two ends of the low-density texture are designated as the low end, and the two ends of the high-density texture are designated as the high end. Select one of the low ends as the target low end, select the low end that is closest to the target low end as the feature low end, and select the high end that is closest to the target low end as the feature high end. If the distance from the bottom of the target to the bottom of the feature is less than the distance from the bottom of the target to the top of the feature, then connect the aggregated line segment between the bottom of the target and the bottom of the feature; otherwise, do not perform any processing. When the target low end traverses all low ends, the low-density texture and the texture formed by the connected low-density texture and the aggregated line segments are regarded as low-aggregate texture. Select one of the high-ends as the target high-end, the high-end closest to the target high-end as the benchmark high-end, and the low-end closest to the target high-end as the benchmark low-end; If the distance from the target high end to the base high end is less than the distance from the target high end to the base low end, then connect the convergent line segment between the target high end and the base high end; otherwise, do not perform any processing. When the target high-end traverses all the high-ends, the high-density texture and the texture formed by the aggregated line segments connected into one are called high-aggregate texture.
[0029] Taking settlement as an example, slip is similar. Ideally, the settlement area should be surrounded by a low-density ridge. However, when the low-density ridge is formed, there may be gaps in the middle because the area around the settlement area may not have settled completely. Therefore, the settlement area may be surrounded by several broken low-density ridges. Since it does not form a closed loop, it needs to be completed. The low-density ridges and high-density ridges that form the slip area also have the same problem, so they also need to be completed. When completing the low-density textures that enclose the settlement area, it is easy to know that the distance from the target low end to the feature low end is less than the distance from the target low end to the feature high end. Therefore, by connecting the convergent line segments between the target low end and the feature low end, the low-density textures that enclose the settlement area can be completed into a loop, thereby forming the settlement area. When completing the low-density and high-density textures that form the slip zone, the low-density textures in the slip zone need to be connected as one unit, and the high-density textures in the slip zone need to be connected as one unit. At the two lower ends of the slip zone, the low-density and high-density textures are adjacent. It is easy to see that the distance from the target low end to the feature low end is greater than the distance from the target low end to the feature high end. Therefore, no processing is required. At the same time, the distance from the target high end to the reference high end is greater than the distance from the target high end to the reference low end. Therefore, no processing is required. However, the completion operation is performed on all positions in the slip zone except for the two lower ends. In this way, the low-density and high-density textures that form the slip zone can be connected as one unit. Thus, high-polymer textures and low-polymer textures can be used to form the settlement zone and the slip zone in the future.
[0030] Reference Figure 8 As shown, matching high-polymerization patterns with low-polymerization patterns and classifying low-polymerization patterns into slippage patterns and settling patterns includes the following steps: Match the high-polymer texture to the low-polymer texture that contacts both ends of it. If there is no low-polymer texture that contacts both ends of the high-polymer texture, match the low-polymer texture with the smallest sum of distances to both ends of the high-polymer texture to the high-polymer texture. Low-polymer patterns that match high-polymer patterns are designated as slip patterns, while low-polymer patterns that do not match high-polymer patterns are designated as settling patterns.
[0031] When turf is moved, insufficient soil may not be removed in some areas, which may result in gaps between high-polymer patterns and their corresponding low-polymer patterns. In this case, since the high-polymer patterns and their corresponding low-polymer patterns form a slippage area, the sum of the distances between the left and right ends of the high-polymer patterns and their corresponding low-polymer patterns is minimized. Therefore, high-polymer patterns and low-polymer patterns are matched accordingly.
[0032] Reference Figure 9 As shown, forming at least one settlement zone and at least one slip zone includes the following steps: The area enclosed by the settlement lines is defined as the settlement area; If a high-polymerization texture and its matching slip texture form a closed loop, then the area enclosed by the high-polymerization texture and its matching slip texture is taken as the local area, and the area covered by the local area and the slip texture matched by the high-polymerization texture is taken as the slip area. If the high-polymer texture and its matching slip texture do not form a closed loop, then connect the left end of the high-polymer texture and its matching slip texture to obtain the first line segment, and connect the right end of the high-polymer texture and its matching slip texture to obtain the second line segment. The area enclosed by the first line segment, the second line segment, the high-polymer texture, and the slip texture matched by the high-polymer texture is defined as the local area, and the area covered by the local area and the slip texture matched by the high-polymer texture is defined as the slip area.
[0033] Since the settlement patterns that form the settlement area are lines, it doesn't matter much whether the settlement area includes the settlement patterns or not. However, the slip zone is created by the downward movement of the turf and includes the moved turf and the soil exposed after the turf moves. The area enclosed by the high-polymer pattern and its matching slip zone only includes the moved turf and does not include the soil exposed after the turf moves. The soil exposed after the turf moves is included by the slip zone matched by the high-polymer pattern and is located above the slip zone matched by the high-polymer pattern. Therefore, the slip zone matched by the high-polymer pattern also needs to be taken into consideration.
[0034] Furthermore, this solution also proposes a storage medium on which a computer-readable program is stored. When the computer-readable program is invoked, it executes the aforementioned method for identifying riverbank slope slippage and settlement defects based on image composite analysis.
[0035] It is understandable that the storage medium can be a magnetic medium, such as a floppy disk, hard disk, or magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid-state drive (SSD).
[0036] In summary, the advantages of this invention are as follows: by eliminating interference in the slope protection image, obtaining the turf density range, forming low-polymer textures, forming high-polymer textures, and forming settlement and slippage areas, aerial photography can be used to acquire images of riverbank slopes. By eliminating interference from obstacles, an interference-free slope protection image is obtained, ensuring the accuracy of identification. During identification, the actual characteristics of slippage and settlement are distinguished, thereby identifying the location of slippage and settlement separately, allowing for targeted repairs, and filling in the gaps in the textures that form slippage and settlement, thus accurately identifying the range of slippage and settlement. Because it uses image visual recognition, the identification efficiency can be greatly improved.
[0037] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.
Claims
1. A method for identifying riverbank revetment slippage and settlement defects based on image composite analysis, characterized in that, include: Obtain images of riverbank revetments covered with turf, remove interference from the revetment images, and obtain surface-processed images. At least one normal region is identified in the surface-processed image, and the turf density range is analyzed within the normal region. Based on the turf density range, at least one low-density block and one high-density block are obtained in the surface-processed image. The low-density block and the high-density block are aggregated respectively to form at least one low-density texture and one high-density texture. Low-density textures and high-density textures are respectively completed and polymerized to form low-polymer textures and high-polymer textures; High-polymer textures were matched with low-polymer textures, and low-polymer textures were classified into slip textures and settling textures. Based on the settlement patterns and slip patterns, at least one settlement area and at least one slip area are formed.
2. The method for identifying riverbank landslide and settlement defects based on image composite analysis according to claim 1, characterized in that, The process of obtaining images of riverbank revetments covered with turf includes the following steps: Select at least one marker point on the riverbank revetment and use the plane perpendicular to the riverbank revetment that passes through the marker point as the reference plane; At least one reference plane divides the riverbank protection into at least one local protection, and the local protection with the largest area is designated as the characteristic local protection. Take a shot perpendicular to the center of the feature slope protection, increase the distance between the camera and the feature slope protection until the feature slope protection is completely contained within the camera's field of view, and use the current distance between the camera and the feature slope protection as the preset distance; Take a picture perpendicular to the center of the local slope protection at a preset distance to obtain an overall image of the local slope protection. Delete the parts other than the local slope protection from the overall image of the local slope protection to obtain a reference image of the local slope protection. The reference images of the partial slope protection are stitched together sequentially to obtain the slope protection image of the river channel.
3. The method for identifying riverbank landslide and settlement defects based on image composite analysis according to claim 2, characterized in that, The process of removing interference from the slope protection image to obtain a surface-processed image includes the following steps: The pixel values of the slope protection soil in the image are obtained in advance as the first pixel value, and the pixel values of the slope protection turf in the image are obtained in advance as the second pixel value; Pixels in the slope protection image whose pixel value is not equal to the first pixel value or the second pixel value are taken as points to be replaced; Adjacent points to be replaced are aggregated to obtain at least one area to be replaced. Feature lines perpendicular to the bottom of the slope protection image are generated on both sides of the area to be replaced. The feature line is shifted to a position tangent to the region to be replaced, and the distance between the two feature lines is taken as the feature distance. The area to be replaced is shifted to the left by a feature distance to obtain a reference area. The image covered by the reference area is used to replace the image in the area to be replaced. After the slope protection image is replaced, the surface-processed image is obtained.
4. The method for identifying riverbank revetment slippage and settlement defects based on image composite analysis according to claim 3, characterized in that, The process of identifying at least one normal region in a surface-processed image includes the following steps: Based on historical data of riverbank protection, the locations where slippage or settlement has occurred in the riverbank protection are identified as abnormal areas. The parts of the surface-processed image other than the abnormal areas are identified as normal parts. The normal parts are then uniformly segmented to obtain at least one normal area.
5. The method for identifying riverbank landslide and settlement defects based on image composite analysis according to claim 4, characterized in that, The process of analyzing the turf density range in a normal area includes the following steps: In the normal area, count the number of pixels whose pixel value is equal to the first pixel value, and obtain the first number; In the normal area, count the number of pixels whose pixel value is equal to the second pixel value to obtain the second number; The first number is superimposed with the second number to obtain the third number. The second number is divided by the third number to obtain the characteristic value of the normal area. The maximum and minimum values of the characteristic value of the normal area are used as endpoints to form the turf density range.
6. The method for identifying riverbank landslide and settlement defects based on image composite analysis according to claim 5, characterized in that, The analysis to obtain at least one low-density block and one high-density block in the surface-processed image includes the following steps: The surface-processed image is uniformly segmented to obtain at least one distribution block. The number of pixels in the distribution block whose pixel value is equal to the second pixel value is counted and used as the fourth number. Divide the fourth number by the total number of pixels in the distribution block to obtain the distribution value of the distribution block; If the distribution value of a distribution area is less than the value at the left end of the turf density range, the distribution area is a low-density area; if the distribution value of a distribution area is greater than the value at the right end of the turf density range, the distribution area is a high-density area.
7. The method for identifying riverbank landslide and settlement defects based on image composite analysis according to claim 6, characterized in that, The process of aggregating low-density blocks and high-density blocks to form at least one low-density texture and one high-density texture includes the following steps: If two low-density blocks are in contact, then the two low-density blocks are merged; if two high-density blocks are in contact, then the two high-density blocks are merged. The texture formed by merging low-density blocks is designated as low-density texture, and the texture formed by merging high-density blocks is designated as high-density texture.
8. The method for identifying riverbank landslide and settlement defects based on image composite analysis according to claim 7, characterized in that, The formation of low-polymer textures and the formation of high-polymer textures include the following steps: The two ends of the low-density texture are designated as the low end, and the two ends of the high-density texture are designated as the high end. Select one of the low ends as the target low end, select the low end that is closest to the target low end as the feature low end, and select the high end that is closest to the target low end as the feature high end. If the distance from the bottom of the target to the bottom of the feature is less than the distance from the bottom of the target to the top of the feature, then connect the aggregated line segment between the bottom of the target and the bottom of the feature; otherwise, do not perform any processing. When the target low end traverses all low ends, the low-density texture and the texture formed by the connected low-density texture and the aggregated line segments are regarded as low-aggregate texture. Select one of the high-ends as the target high-end, the high-end closest to the target high-end as the benchmark high-end, and the low-end closest to the target high-end as the benchmark low-end; If the distance from the target high end to the base high end is less than the distance from the target high end to the base low end, then connect the convergent line segment between the target high end and the base high end; otherwise, do not perform any processing. When the target high-end traverses all high-ends, the high-density texture and the texture formed by the aggregated line segments connected as a whole are called high-aggregation texture.
9. The method for identifying riverbank landslide settlement defects based on image composite analysis according to claim 8, characterized in that, The process of matching high-polymerization patterns with low-polymerization patterns and classifying low-polymerization patterns into slippage patterns and settling patterns includes the following steps: Match the high-polymer texture to the low-polymer texture that contacts both ends of it. If there is no low-polymer texture that contacts both ends of the high-polymer texture, match the low-polymer texture with the smallest sum of distances to both ends of the high-polymer texture to the high-polymer texture. Low-polymer patterns that match high-polymer patterns are designated as slip patterns, while low-polymer patterns that do not match high-polymer patterns are designated as settling patterns.
10. The method for identifying riverbank revetment slippage and settlement defects based on image composite analysis according to claim 9, characterized in that, The formation of at least one settlement zone and at least one slip zone includes the following steps: The area enclosed by the settlement lines is defined as the settlement area; If a high-polymerization texture and its matching slip texture form a closed loop, then the area enclosed by the high-polymerization texture and its matching slip texture is taken as the local area, and the area covered by the local area and the slip texture matched by the high-polymerization texture is taken as the slip area. If the high-polymer texture and its matching slip texture do not form a closed loop, then connect the left end of the high-polymer texture and its matching slip texture to obtain the first line segment, and connect the right end of the high-polymer texture and its matching slip texture to obtain the second line segment. The area enclosed by the first line segment, the second line segment, the high-polymer texture, and the slip texture matched by the high-polymer texture is defined as the local area, and the area covered by the local area and the slip texture matched by the high-polymer texture is defined as the slip area.
