Integrated slope monitoring system and method based on laser radar

Abstract

The invention discloses an integrated side slope monitoring system and method based on a laser radar, and relates to the technical field of side slope monitoring, and the method comprises the following steps: setting the laser radar and a high-definition camera for a side slope under the jurisdiction of a current region, carrying out the three-dimensional scanning of the side slope through the laser radar, and obtaining the point cloud data of the surface of the side slope, based on the soil characteristics of the current region, grouping and dividing the jurisdiction slopes, and recording the collected slope data; a slope abnormal site is determined through abnormal point cloud density and elevation values in point cloud data in the slope, coordinates of the abnormal site are obtained through position coordinate conversion, and main side view image acquisition is carried out on the coordinate position based on a high-definition camera; and performing line extraction on the main view and the side view of the abnormal position, and performing parameter analysis on line curvature and distance in the main view and the side view. According to the method, the advantages of laser radar and image analysis are combined, the data processing amount in the slope monitoring process is reduced, and the anomaly recognition accuracy and the monitoring efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of slope monitoring. Specifically, it relates to an integrated slope monitoring system and method based on lidar. Background Technique

[0002] Slope stability monitoring is an important safety guarantee link in fields such as geological engineering, transportation construction, and mine exploitation. Traditional slope monitoring methods, such as manual inspections and total station measurements, have problems such as low efficiency, strong subjectivity, and low monitoring frequency. It is difficult to capture the subtle changes of slopes in real time and cannot meet the requirements of modern engineering for the timeliness and accuracy of safety monitoring. With the development of technology, although the monitoring technology based on image recognition can intuitively reflect the surface conditions of slopes, in the face of large-area slopes, it is necessary to process a large amount of image data, with a large computational load and low analysis efficiency, resulting in relatively high false positive and false negative rates in anomaly recognition. Existing lidar monitoring technologies, although able to quickly obtain the three-dimensional point cloud data of slopes, lack in-depth analysis and accurate judgment of abnormal data, and it is difficult to accurately distinguish real dangerous abnormal situations from complex point cloud data, nor can they effectively predict potential slope abnormal areas, having problems of low practicality and functionality.

[0003] Regarding the problems in the related technology, no effective solution has been proposed yet. Summary of the Invention

[0004] Regarding the problems in the related technology, the present invention proposes an integrated slope monitoring system and method based on lidar to overcome the above-mentioned technical problems existing in the existing related technology.

[0005] To this end, the specific technical solution adopted by the present invention is as follows: An integrated slope monitoring method based on lidar, the method includes the following steps: S1. Set lidars and high-definition cameras for the slopes under the jurisdiction of the current area respectively. Perform three-dimensional scanning on the slopes through the lidars to obtain the surface point cloud data of the slopes. Based on the soil characteristics of the current area, group and divide the slopes under the jurisdiction and record the collected slope data. S2. Determine the abnormal sites of the slopes through the abnormal point cloud density and elevation values in the point cloud data of the slopes. Obtain the coordinates of the abnormal sites through position coordinate conversion, and perform main and side view image acquisition on the coordinate positions based on the high-definition cameras. S3. Extract the lines from the main and side views of the abnormal positions, perform parameter analysis on the line curvatures and distances in the main and side views, determine the types of the abnormal point positions of the slopes, including depressions, protrusions, and cracks, and combine the remaining same abnormal parameters under the same soil characteristics in the same group to determine the abnormalities of the slopes. S4. Image acquisition of the slope is carried out based on a high-definition camera. Manually mark potential abnormal positions, extract data of potential abnormal positions and reconcile them with lidar point cloud data. Based on the data upload results of potential abnormal positions, monitor and identify potential abnormalities of the slope.

[0006] As a preferred implementation, S1 includes the following steps: S11. Based on the area of the subordinate slope and the parameters of using lidar, set up lidar on the slope, and scan the slope through the lidar to obtain surface point cloud data; S12. Collect geological exploration reports and soil samples of different slopes in the current area, including soil type, density, water content, and particle size distribution. Determine slopes with the same soil type and similar density, water content, and particle size distribution within a similar error range as the same type of slopes. By establish a slope monitoring database, divide the same type of slopes into the same group, and further subdivide them into layered structure slopes, block structure slopes, and reticular structure slopes according to the rock stratum structure of the slope. At the same time, record the numbers of the slopes and the point cloud data.

[0007] As a preferred implementation, S2 includes the following steps: S21. Divide different regions of the subordinate slope, and mark abnormal regions according to the point cloud density and elevation value in each region, combined with the normal values of the point cloud density and elevation value of the current slope; S22. For the point cloud data obtained by scanning each slope, establish a three-dimensional space coordinate system and mark the three-dimensional center coordinates of each divided region to obtain the characteristic coordinates of the marked abnormal region. Based on the characteristic coordinates of the abnormal region, collect the front view image and side view image of the abnormal site in the abnormal region corresponding to the characteristic coordinates through a high-definition camera, and upload them to the slope monitoring database.

[0008] As a preferred implementation, S21 includes the following steps: S211. For the point cloud data obtained from the slope, divide the point cloud data on the slope into grids at a fixed interval to obtain regions, count the number of point clouds in each region as the point cloud density feature of the current region, select the region on the current slope without signs of abnormal deformation or damage as the normal sample region to calculate the mean and standard deviation of the point cloud density feature, and take the range from mean - 2 standard deviations to mean + 2 standard deviations as the normal interval, and determine the region outside the normal interval as the abnormal point cloud density region; S212. At the same time, for each region in the current regions of the point cloud data points , based on the radius determine the neighborhood of the current point cloud data. For all points within the neighborhood of the point, calculate the average value of their elevation values respectively , where represents the points in the neighborhood, represents the elevation value of the -th point in the neighborhood. Calculate the elevation value of the current point and the difference from the average elevation value of the neighborhood . When , it means that there is an abnormal elevation value in the current area, and the current area is determined as an area with abnormal elevation value, where represents the elevation value threshold. represents the elevation value threshold.

[0009] As a preferred embodiment, S3 includes the following steps: S31. According to the front view image and side view image of the obtained abnormal area, perform two-dimensional conversion and line extraction on the images, and perform parameter analysis on the line curvature and interval distance in the front view image and side view image to determine the slope abnormal type of the current abnormal area; S32. Through the slope grouping situation in the slope monitoring database, statistically analyze the slope abnormalities under the subdivision items in the same group according to the type, and analyze and determine the existing serious slope abnormalities.

[0010] As a preferred embodiment, S31 includes the following steps: S311. For the front view image and side view image of the obtained abnormal area, perform grayscale processing on the images through the formula , where R, G, and B respectively represent the pixel values of the red, green, and blue channels. At the same time, perform histogram equalization on the images to enhance the contrast of the images; S312. Use the edge detection algorithm to extract the abnormal edge information in the abnormal area image, use the Gaussian filter to smooth the image to reduce the influence of noise, calculate the gradients of the image in the horizontal and vertical directions to obtain the gradient amplitude and gradient direction, perform non-maximum suppression in the gradient direction to retain the points with the largest local gradient amplitude to refine the edge, and at the same time mark the edge points through double-threshold detection. The steps are as follows: Set the thresholds , , where < . Mark the points with gradient amplitude greater than as strong edge points, mark the points with gradient amplitude less than as non-edge points, and mark the points with gradient amplitude between , as pending points. When a pending point is connected to a strong edge point, it is determined as a strong edge point, and when a pending point is connected to a non-edge point, it is determined as a non-edge point; S313. Perform line fitting on the obtained strong edge points. Use the polynomial fitting method to determine the line parameters by minimizing the sum of squared errors. Obtain the line curvature by calculating the curvature of the line, discretize the line to calculate the discrete curvature, and select three points on the curve to calculate the discrete curvature using the reciprocal of the circumradius of the triangle. Meanwhile, select multiple points to calculate the vertical distance between corresponding points of two adjacent lines. Take the average value as the current line spacing distance. ; S314. Based on the discrete curvature and line spacing of the lines in different abnormal regions obtained, through the spacing distance threshold and the curvature threshold , determine the abnormal type: When among the lines obtained in the abnormal region and , it represents that the lines in the current region show a trend of sagging inward towards the slope, and it is determined as a sagging abnormality; When the lines obtained in the abnormal region only meet , it represents that there is a crack abnormality in the current region; When the lines obtained in the abnormal region and , it represents that the curvature of the line shows a trend of bulging outward towards the slope, and the line spacing distance in the bulging region has no obvious change, and it is determined as a bulging abnormality; When the above three situations are not met, mark the current region as normal.

[0011] As a preferred implementation manner, S32 includes the following steps: S321. Based on the slope numbers in the sub-items under the same grouping in the slope monitoring database, respectively extract the sagging, bulging, and crack abnormal data, including curvature and spacing distance. Determine the severe outliers of the curvature and spacing distance in the sub-items through the interquartile range method. The specific steps are as follows: For the sagging, bulging, and crack abnormal data, respectively obtain the curvature and spacing distance values of the slopes in the same sub-item under the corresponding abnormality. Sort the data from small to large, and calculate the first quartile and the third quartile , calculate the interquartile range , where represents the interquartile range; Obtain the lower limit and the upper limit , determine the curvature and spacing distance values of the slopes in the same sub-item under the same abnormality. Mark the abnormalities in the data that are less than and greater than as severe abnormalities.

[0012] As a preferred embodiment, S4 includes the following steps: S41. Collect images of the slope at fixed intervals through a high-definition camera, identify and mark potential abnormal positions manually, and determine the position coordinates of the manually marked potential abnormal positions through the point cloud data of the lidar; S42. Compare the coordinates of the potential abnormal positions with the characteristic coordinates of the marked abnormal areas uploaded in the slope monitoring database. When the coordinates of the potential abnormal positions match the characteristic coordinates of the existing marked abnormal areas, the current potential abnormal positions are ignored; S43. When the coordinates of the potential abnormal positions do not match the characteristic coordinates of the existing marked abnormal areas, repeat S3 to analyze and determine the potential abnormalities.

[0013] As a preferred embodiment, the following steps are further included in S43: S431. When the coordinates of the potential abnormal positions do not match the characteristic coordinates of the existing marked abnormal areas, simultaneously extract the point cloud data within the coordinate area of the current potential abnormal position, calculate the similarity of the point cloud data between other areas and the coordinate area of the current abnormal position through the Euclidean distance, mark similar abnormalities for the areas with a similarity less than 0.3, and repeat S3 to analyze and determine the abnormalities in the similar abnormal areas.

[0014] An integrated slope monitoring system based on lidar, comprising a data acquisition module, a parameter conversion and abnormality recognition module, and a potential abnormality determination module: The data acquisition module includes a lidar and a high-definition camera. The three-dimensional scan of the slope is performed through the lidar to obtain the point cloud data of the slope surface, the image data of the abnormal positions of the slope is collected through the high-definition camera, and at the same time, based on the soil characteristics of the current area, through A slope monitoring database is established to group and divide the subordinate slopes, and the collected slope data is recorded; The parameter conversion and abnormality recognition module determines the abnormal sites of the slope through the abnormal point cloud density and elevation value in the point cloud data of the slope, obtains the coordinates of the abnormal sites through position coordinate conversion, performs front and side view image acquisition of the coordinate positions based on the high-definition camera, extracts the lines of the front and side views of the abnormal positions, analyzes the parameters of the line curvature and distance in the front and side views, determines the types of the abnormal sites of the slope, including depressions, protrusions, and cracks, and combines the remaining same abnormal parameters under the same soil characteristics in the same group to determine the abnormalities of the slope; The potential anomaly determination module marks potential anomaly positions manually from the slope images collected by a high-definition camera, extracts the data of potential anomaly positions and reconciles them with the lidar point cloud data, and monitors and identifies potential slope anomalies and similar anomalies based on the data upload results of the potential anomaly positions.

[0015] The beneficial effects of the present invention are as follows: 1. The present invention collects point cloud data of the slope by lidar to determine the anomaly position, further obtains the image of the anomaly position, extracts the lines from the image to highlight the structural features of the anomaly, reduces the computational amount of anomaly recognition during the monitoring process, and at the same time analyzes according to the same anomaly data under the same geological features to determine the serious anomalies among the slope anomalies during the slope monitoring process, facilitating the management personnel to maintain the slope in a timely manner; 2. The present invention first uses lidar to collect point cloud data to determine the anomaly position, then specifically obtains the image of the anomaly position and extracts the lines, which can focus on the key areas and features. Compared with the long-term comprehensive monitoring and analysis of the entire slope image, this method greatly reduces the amount of data to be processed, thus significantly shortening the calculation time, improving the operation efficiency, obtaining the monitoring results faster, enhancing the practicability, and at the same time extracting the lines from the image of the anomaly position can highlight the structural features of the anomaly, such as cracks, depressions, protrusions, and these key anomaly features are clearer after the line extraction, reducing the probability of misjudgment and missed judgment during the monitoring process; 3. The present invention obtains the overall image of the monitored slope regularly, manually marks the potential anomalies, combines the potential anomaly coordinates and the feature coordinates of the existing anomaly areas, analyzes the potential anomalies in the unrecorded areas, and at the same time determines the similarity of the point cloud data in the potential anomaly areas, and synchronously analyzes the similar anomalies in the remaining areas of the slope that meet the similarity determination, further improving the accuracy of slope monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings without creative efforts based on these drawings.

[0017] Figure 1 is a flowchart of an integrated slope monitoring method based on lidar according to an embodiment of the present invention; Figure 2 is a block diagram of an integrated slope monitoring system based on lidar according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] To further illustrate the embodiments, the present invention provides accompanying drawings, which are part of the disclosure of the present invention. These drawings are mainly used to illustrate the embodiments and can be combined with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, those of ordinary skill in the art should be able to understand other possible implementation manners and the advantages of the present invention. The components in the drawings are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0019] According to an embodiment of the present invention, an integrated slope monitoring system and method based on lidar are provided.

[0020] The present invention will be further described below in conjunction with the accompanying drawings and specific implementation manners: Embodiment 1: As Figure 1 shown, an integrated slope monitoring method based on lidar according to an embodiment of the present invention includes the following steps: S1. Respectively set lidars and high-definition cameras for the slopes under the jurisdiction of the current area. Perform three-dimensional scanning on the slopes through the lidars to obtain slope surface point cloud data. Based on the soil characteristics of the current area, group and divide the slopes under the jurisdiction, and record the collected slope data; S11. Based on the area of the slopes under the jurisdiction and the parameters of the lidars used, set the lidars on the slopes, and perform scanning on the slopes through the lidars to obtain surface point cloud data; S12. Collect geological exploration reports and soil samples of different slopes in the current area, including soil types, densities, water contents, and particle gradations. Determine the slopes with the same soil type and similar densities, water contents, and particle gradations within a similar error range as the same type of slopes. By establishing a slope monitoring database, divide the same type of slopes into the same group. At the same time, according to the rock stratum structure of the slopes, further subdivide them into layered structure slopes, massive structure slopes, and reticular structure slopes in the group, and record the numbers and point cloud data of the slopes.

[0021] It should be noted that in the process of determining the consistency of soil types, according to the domestic general soil classification standard, clarify the soil types to which the soils of each slope belong. If the soils of two slopes belong to the same classification unit, such as both being silty clay, then the two are considered to have the same soil type. Measure the densities of the soils of each slope by the cutting ring method, wax sealing method, etc., calculate the difference in soil densities of different slopes and judge through the error range, where the soil density error range is ±0.5 g / cm³. Measure the water content of the soils of each slope by the drying method, and compare the average water content of the soils of different slopes. If the difference is within ±3%, it indicates that the water content of the current slope is similar. Select the effective particle size, average particle size, and restricted particle size of different slopes, calculate the ratio of the soil characteristic particle sizes of different slopes. When the ratios of the characteristic particle sizes are all between 0.9 - 1.1, it represents that the particle gradations of the current slopes are similar.

[0022] S2. Determine the abnormal site of the slope through the abnormal point cloud density and elevation value in the midpoint cloud data of the slope, obtain the coordinates of the abnormal site through position coordinate conversion, and collect the main side view images of the coordinate position based on the high-definition camera; S21. Divide the subordinate slopes into different regions, and mark the abnormal regions according to the point cloud density and elevation value in each region, combined with the normal values of the point cloud density and elevation value of the current slope; S211. For the point cloud data obtained for the slope, divide the point cloud data on the slope into grids at a fixed interval to obtain regions, count the number of point clouds in each region as the point cloud density feature of the current region, select the region on the current slope without signs of abnormal deformation or damage as the normal sample region to calculate the mean and standard deviation of the point cloud density feature, and use the range from mean - 2 standard deviations to mean + 2 standard deviations as the normal interval, and determine the region outside the normal interval as the abnormal point cloud density region; It should be noted that when determining the region on the current slope without signs of abnormal deformation or damage, it is necessary to determine by combining the on-site photos of the slope and consulting experts in the current field and relying on the empirical method.

[0023] S212. At the same time, for each region in the current regions of the point cloud data points , based on the radius determine the neighborhood of the current point cloud data. For all points within the neighborhood of the point, calculate the average value of their elevation values , where represents the points within the neighborhood, represents the elevation value of the th point within the neighborhood, calculate the difference between the elevation value of the current point and the neighborhood average elevation value . When , it means that there is an abnormal elevation value in the current region, and the current region is determined as an abnormal elevation value region, where represents the elevation value threshold.

[0024] It should be noted that the radius needs to be determined based on the density of the point cloud data in the current region by the empirical method, and the threshold is determined by consulting experts in the current field by combining factors such as the terrain complexity and data accuracy of the current region to determine an appropriate threshold.

[0025] S22. For the point cloud data obtained by scanning each slope, establish a three-dimensional space coordinate system and mark the three-dimensional central coordinates of each divided area to obtain the characteristic coordinates for marking the abnormal area. Based on the characteristic coordinates of the abnormal area, collect the front view image and side view image of the abnormal site in the abnormal area corresponding to the characteristic coordinates through a high-definition camera, and upload them to the slope monitoring database.

[0026] Embodiment 2: S3. Extract the lines from the front view and side view of the abnormal position, analyze the parameters of the line curvature and distance in the front view and side view, and determine the types of abnormal points on the slope, including depressions, protrusions, and cracks. Combine the remaining same abnormal parameters under the same soil characteristics in the same group to determine the abnormality of the slope. S31. According to the front view image and side view image of the obtained abnormal area, perform two-dimensional conversion and line extraction on the images, and analyze the parameters of the line curvature and interval distance in the front view image and side view image to determine the slope abnormal type of the current abnormal area. S311. For the front view image and side view image of the obtained abnormal area, perform grayscale processing on the images through the formula where R, G, and B represent the pixel values of the red, green, and blue channels respectively. At the same time, perform histogram equalization on the images to enhance the contrast of the images. It should be noted that when adjusting the histogram of the image, assume that the gray value of the original image is , and the transformed gray value is , then the transformation function of histogram equalization is: ; Among them, is the total number of gray levels, represents the gray level 's probability density function.

[0027] S312. Use the edge detection algorithm to extract the abnormal edge information in the abnormal area image, use the Gaussian filter to smooth the image to reduce the influence of noise, calculate the gradients of the image in the horizontal and vertical directions, obtain the gradient amplitude and gradient direction, perform non-maximum suppression in the gradient direction to retain the points with the largest local gradient amplitude to refine the edge, and at the same time mark the edge points through double-threshold detection. The steps are as follows: Set the thresholds , , where < , mark the points with gradient amplitude greater than as strong edge points, mark the points with gradient amplitude less than as non-edge points, and mark the points with gradient amplitude between , The point between them is marked as the to-be-determined point. When the to-be-determined point is connected to a strong edge point, it is determined as a strong edge point; when the to-be-determined point is connected to a non-edge point, it is determined as a non-edge point. It should be noted that through double-threshold processing, false edges caused by noise can be effectively removed, while the true abnormal edges in the image are retained. The settings of the high threshold and the low threshold can be determined by consulting experts in related fields and combining empirical methods according to the actual usage situation.

[0028] S313. Perform line fitting on the obtained strong edge points. The polynomial fitting method is used to determine the line parameters by minimizing the sum of squared errors. Calculate the line curvature to obtain the line curvature degree. Discretize the line to calculate the discrete curvature. Select three points on the curve to calculate the discrete curvature through the reciprocal of the circumradius of the triangle. At the same time, select to calculate the vertical distance between corresponding points of two adjacent lines at multiple points. Take the average value as the current line interval distance. ; S314. Through the discrete curvature and line spacing of the lines in different abnormal regions obtained, through the interval distance threshold and the curvature threshold , determine the abnormal type: When among the lines obtained in the abnormal region and , it represents that the lines in the current region show a trend of concaving inward towards the slope, and it is determined as a concave anomaly; When the lines obtained in the abnormal region only meet , it represents that there is a crack anomaly in the current region; When the lines obtained in the abnormal region and , it represents that the curvature of the line shows a trend of bulging outward towards the slope, and the line interval distance in the bulging region has no obvious change, and it is determined as a bulge anomaly; When the above three situations are not met, mark the current region as no anomaly.

[0029] It should be noted that the interval distance threshold and the curvature threshold need to collect a large amount of slope monitoring image data similar to the current slope geological conditions, environmental factors, and data acquisition methods, and label the normal regions and known abnormal regions in these data. Calculate the curvature and interval distance of the lines for the normal regions and abnormal regions to obtain a series of curvature values and interval distance values. Build a training set according to the labels of abnormal types to build a machine learning model, and determine the interval distance threshold and curvature threshold by analyzing the decision boundary of the model.

[0030] S32. Based on the grouped slopes under the slope monitoring database, count the slope anomalies in the same group for the sub-items according to their types, and analyze and determine the existing serious slope anomalies. S321. Based on the slope numbers in the sub-items under the same group in the slope monitoring database, extract the depression, bulge, and crack anomaly data respectively, including curvature and spacing distance. Determine the serious outliers of the curvature and spacing distance in the sub-items by using the interquartile range method. The specific steps are as follows: For the depression, bulge, and crack anomaly data, obtain the curvature and spacing distance values of the slopes in the same sub-item corresponding to the respective anomalies. Sort the data from smallest to largest, and calculate the first quartile and the third quartile , calculate the interquartile range , where represents the interquartile range; Obtain the lower limit and the upper limit . Determine the curvature and spacing distance values of the slopes in the same sub-item under the same anomaly. Mark the anomalies in the data that are less than and greater than as serious anomalies.

[0031] It should be noted that by jointly analyzing the anomalies with similar geological types and determining the distribution range of the data through the quartiles of the data, the data outside this range in the same sub-item are serious anomalies, which is convenient for subsequent maintenance of the slopes with serious anomalies and further observation of the anomalies outside the range.

[0032] S4. Collect images of the slope based on a high-definition camera, manually mark the potential anomaly positions, extract the data of the potential anomaly positions and reconcile them with the lidar point cloud data, and monitor and identify the potential slope anomalies based on the data upload results of the potential anomaly positions. S41. Collect images of the slope at fixed intervals based on a high-definition camera, manually identify and mark the potential anomaly positions, and determine the position coordinates of the manually marked potential anomaly positions through the lidar point cloud data. S42. Compare the coordinates of the potential anomaly positions with the characteristic coordinates of the marked anomaly areas uploaded in the slope monitoring database. When the coordinates of the potential anomaly positions match the characteristic coordinates of the existing marked anomaly areas, ignore the current potential anomaly positions. S43. When the coordinates of the potential anomaly positions do not match the characteristic coordinates of the existing marked anomaly areas, repeat S3 to analyze and determine the potential anomalies.

[0033] It should be noted that the slope is imaged by a high-definition camera at fixed intervals with a period of 7 days, which can be shortened or extended according to the actual situation.

[0034] S431. When the coordinates of the potential abnormal position do not match the characteristic coordinates of the existing marked abnormal area, the point cloud data within the current potential abnormal position coordinate area is extracted simultaneously. The similarity between the point cloud data in other areas and that in the current abnormal position coordinate area is calculated through the Euclidean distance. Similar abnormal markings are made for the areas with a similarity less than 0.3, and S3 is repeated to analyze and determine the abnormalities in the similar abnormal areas.

[0035] It should be noted that during the process of identifying some tiny cracks, the point cloud resolution of the lidar may not meet the requirements for identifying these tiny cracks. Tiny cracks may only appear as abnormalities of a few points in the point cloud map, making it difficult to distinguish them from the surrounding noise or normal terrain undulations and easy to be ignored. By manually annotating the potential abnormal areas and combining with the lidar for coordinate determination, the potential abnormalities and similar abnormalities can be further identified.

[0036] Embodiment 3: As Figure 2 shown, an integrated slope monitoring system based on lidar includes a data acquisition module, a parameter conversion and anomaly recognition module, and a potential anomaly determination module: The data acquisition module includes a lidar and a high-definition camera. The three-dimensional scan of the slope is carried out by the lidar to obtain the point cloud data of the slope surface, and the image data of the abnormal position of the slope is collected by the high-definition camera. At the same time, based on the soil characteristics of the current area, a slope monitoring database is established to group and divide the subordinate slopes, and the collected slope data is recorded; The parameter conversion and anomaly recognition module determines the abnormal sites of the slope through the abnormal point cloud density and elevation value in the point cloud data of the slope, obtains the coordinates of the abnormal sites through position coordinate conversion, collects the main and side view images of the coordinate positions based on the high-definition camera, extracts the lines in the main and side views of the abnormal positions, analyzes the parameters of the line curvature and distance in the main and side views, determines the types of the abnormal points of the slope, including depression, protrusion, and crack, and combines the remaining same abnormal parameters under the same soil characteristics in the same group to determine the abnormalities of the slope; The potential anomaly determination module marks the potential abnormal positions manually through the slope images collected by the high-definition camera, extracts the data of the potential abnormal positions and reconciles them with the lidar point cloud data, and monitors and identifies the potential abnormalities and similar abnormalities of the slope based on the data upload results of the potential abnormal positions.

[0037] In summary, the present invention collects slope point cloud data through lidar to determine abnormal positions, further acquires images of the abnormal positions, extracts lines from the images to highlight the abnormal structural features, reduces the computational amount of abnormal recognition during the monitoring process, and analyzes the same abnormal data under the same geological features to determine the serious abnormalities among the slope abnormalities during slope monitoring, facilitating the timely maintenance of the slope by management personnel; By first using lidar to collect point cloud data to determine abnormal positions, and then specifically acquiring images of the abnormal positions and extracting lines, it is possible to focus on key areas and features. Compared with the long-term comprehensive monitoring and analysis of the entire slope image, this method greatly reduces the amount of data to be processed, thereby significantly shortening the calculation time, improving the operation efficiency, obtaining monitoring results faster, enhancing the practicality. At the same time, extracting lines from the images of abnormal positions can highlight the abnormal structural features, such as cracks, depressions, and protrusions. These key abnormal features are clearer after line extraction, reducing the probability of misjudgment and missed judgment during the monitoring process.

[0038] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An integrated slope monitoring method based on lidar, characterized in that, The method includes the following steps: S1. Set lidar and high-definition cameras for the slopes under the jurisdiction of the current area respectively. Conduct three-dimensional scanning of the slopes through the lidar to obtain the point cloud data of the slope surface. Based on the soil characteristics of the current area, group and divide the slopes under the jurisdiction, and record the collected slope data; S2. Determine the abnormal sites of the slopes through the abnormal point cloud density and elevation value in the point cloud data of the slopes. Obtain the coordinates of the abnormal sites through position coordinate conversion, and collect the main and side view images of the coordinate positions based on the high-definition cameras; S3. Extract the lines from the main and side views of the abnormal positions, analyze the parameters of the line curvature and distance in the main and side views, determine the types of the abnormal points of the slopes, including depressions, protrusions, and cracks. Combine the remaining same abnormal parameters under the same soil characteristics in the same group to determine the abnormalities of the slopes; S4. Collect images of the slopes based on the high-definition cameras, manually mark the potential abnormal positions, extract the data of the potential abnormal positions and reconcile them with the lidar point cloud data. Based on the data upload results of the potential abnormal positions, monitor and identify the potential abnormalities of the slopes.

2. The integrated slope monitoring method based on lidar according to claim 1, wherein, The S1 includes the following sub-steps: S11. Based on the area of the slopes under the jurisdiction and the parameters of using the lidar, set the lidar on the slopes, and conduct scanning of the slopes through the lidar to obtain the surface point cloud data; S12. Collect the geological exploration reports and soil samples of different slopes in the current area, including soil type, density, water content, and particle size distribution. Determine slopes with the same soil type and similar density, water content, and particle size distribution within a similar error range as the same type of slopes. Through establish a slope monitoring database, divide the same type of slopes into the same group, and further subdivide them into layered structure slopes, massive structure slopes, and reticular structure slopes according to the rock stratum structure of the slopes. At the same time, record the slope numbers and point cloud data.

3. The integrated slope monitoring method based on lidar according to claim 2, wherein, The S2 includes the following steps: S21. Divide different regions for the slopes under the jurisdiction, and mark the abnormal regions according to the point cloud density and elevation value in each region, combined with the normal values of the point cloud density and elevation value of the current slope; S22. For the point cloud data obtained by scanning each slope, establish a three-dimensional space coordinate system and mark the three-dimensional center coordinates of each divided region to obtain the characteristic coordinates of the marked abnormal regions. Based on the characteristic coordinates of the abnormal regions, collect the main view image and side view image of the abnormal sites of the abnormal regions corresponding to the characteristic coordinates through the high-definition cameras, and upload them to the slope monitoring database.

4. The integrated slope monitoring method based on lidar according to claim 3, characterized in that, The S21 includes the following steps: S211. For the point cloud data obtained for the slope, divide the point cloud data on the slope into grids at a fixed interval to obtain regions, count the number of point clouds in each region as the point cloud density feature of the current region, select the regions on the current slope without signs of abnormal deformation or damage as the normal sample regions to calculate the mean and standard deviation of the point cloud density features, take the range from mean - 2 standard deviations to mean + 2 standard deviations as the normal interval, and determine the regions not within the normal interval as regions with abnormal point cloud density; S212. At the same time, for each point cloud data point in the current area, based on the radius to determine the neighborhood of the current point cloud data. For all points within the neighborhood of a point, calculate the average value of their elevation values , where represents the points within the neighborhood, represents the elevation value of the th point within the neighborhood. Calculate the difference between the elevation value of the current point and the average neighborhood elevation value . When , it means that there is an abnormal elevation value in the current area, and the current area is determined as an abnormal elevation value area, where represents the elevation value threshold.

5. The integrated slope monitoring method based on lidar according to claim 2, characterized in that, The S3 includes the following steps: S31. According to the obtained main view image and side view image of the abnormal region, conduct two-dimensional conversion and line extraction on the images, analyze the parameters of the line curvature and interval distance in the main view image and the side view image to determine the slope abnormal type of the current abnormal region; S32. Through the grouping situation of the slopes under the jurisdiction in the slope monitoring database, statistically analyze the slope abnormalities under the same sub-items in the same group according to the types, and analyze and determine the existing serious slope abnormalities.

6. The integrated slope monitoring method based on lidar according to claim 5, wherein, The S31 includes the following steps: S311. For the front view image and side view image of the obtained abnormal area, use the formula to grayscale the image, where R, G, and B represent the pixel values of the red, green, and blue channels respectively. At the same time, perform histogram equalization on the image to enhance the contrast of the image; S312. Use the edge detection algorithm to extract the abnormal edge information in the abnormal region image, use the Gaussian filter to smooth the image to reduce the influence of noise, calculate the gradients of the image in the horizontal and vertical directions to obtain the gradient amplitude and gradient direction, and conduct non-maximum suppression in the gradient direction to retain the points with the largest local gradient amplitude to refine the edge. At the same time, mark the edge points through double-threshold detection, and its steps are as follows: Set a threshold and , where < , mark the points with gradient magnitude greater than as strong edge points, mark the points with gradient magnitude less than as non-edge points, and mark the points with gradient magnitude between and as undetermined points. If an undetermined point is connected to a strong edge point, it is determined as a strong edge point; if an undetermined point is connected to a non-edge point, it is determined as a non-edge point; S313. Perform line fitting on the obtained strong edge points. Use the polynomial fitting method to determine the line parameters by minimizing the sum of squared errors. Obtain the line curvature by calculating the curvature of the line. Discretize the line to calculate the discrete curvature. Select three points on the curve and calculate the discrete curvature using the reciprocal of the circumradius of the triangle. Meanwhile, select multiple points to calculate the vertical distance between corresponding points of two adjacent lines. Take the average value as the current line interval distance. ; S314. Determine the type of anomaly by obtaining the discrete curvature and line spacing of different anomaly area lines, and through the interval distance threshold and the curvature threshold , as follows: When the lines obtained in the abnormal area and it indicates that the lines in the current area tend to be sunken towards the inside of the slope, and it is determined as a sunken abnormality; When the lines obtained in the abnormal area only conform to it means that there is a crack abnormality in the current area; When the lines obtained in the abnormal area and it means that when the curvature of the line shows a trend of bulging outwards towards the slope, and the line spacing in the bulging area has no obvious change, it is determined as a bulging abnormality; When the above three situations are not met, mark the current region as no abnormality.

7. An integrated slope monitoring method based on lidar according to claim 6, characterized in that, The S32 includes the following steps: S321. Based on the slope numbers in the sub-items under the same grouping in the slope monitoring database, respectively extract the depression, protrusion, and crack abnormal data, including curvature and interval distance. Determine the severe outliers of the curvature and interval distance in the sub-items by using the interquartile range method. The specific steps are as follows: For the abnormal data of depressions, protrusions, and cracks, respectively obtain the curvature and interval distance values of the slope in the same sub-item corresponding to the respective abnormalities, sort the data from smallest to largest, and calculate the first quartile and the third quartile , calculate the interquartile range , where represents the interquartile range; Obtain the lower limit and the upper limit , determine the curvature and interval distance values of the slope in the same sub-item under the same anomaly, and mark the anomalies in the data that are less than and greater than as severe anomalies.

8. The integrated slope monitoring method based on lidar according to claim 2, characterized in that, The S4 includes the following steps: S41. Use a high-definition camera to collect images of the slope at fixed intervals. Manually identify and mark potential abnormal positions, and determine the position coordinates of the manually marked potential abnormal positions through the point cloud data of the lidar; S42. Compare the potential abnormal position coordinates with the characteristic coordinates of the marked abnormal areas uploaded in the slope monitoring database. When the potential abnormal position coordinates match the existing characteristic coordinates of the marked abnormal areas, ignore the current potential abnormal position; S43. When the potential abnormal position coordinates do not match the existing characteristic coordinates of the marked abnormal areas, repeat S3 to analyze and determine the potential abnormality.

9. The integrated slope monitoring method based on lidar according to claim 8, wherein, The S43 also includes the following steps: S431. When the potential abnormal position coordinates do not match the existing characteristic coordinates of the marked abnormal areas, simultaneously extract the point cloud data within the area of the current potential abnormal position coordinates. Calculate the similarity of the point cloud data between other areas and the area within the current abnormal position coordinates through the Euclidean distance. Mark similar abnormalities for the areas with a similarity less than 0.3, and repeat S3 to analyze and determine the abnormalities in the similar abnormal areas.

10. An integrated slope monitoring system based on lidar, characterized in that, This system adopts the integrated slope monitoring method based on lidar as described in any one of claims 1-9, including a data acquisition module, a parameter conversion and anomaly recognition module, and a potential anomaly determination module: The data acquisition module includes a lidar and a high-definition camera. The three-dimensional scanning of the slope by the lidar obtains the point cloud data of the slope surface, and the high-definition camera collects the image data of the abnormal position of the slope. At the same time, based on the soil characteristics of the current area, through establish a slope monitoring database to group and divide the subordinate slopes, and record the collected slope data; The parameter conversion and anomaly recognition module determines the abnormal sites of the slope through the abnormal point cloud density and elevation value in the point cloud data of the slope midpoint, obtains the coordinates of the abnormal sites through position coordinate conversion, collects the main and side view images of the coordinate positions based on a high-definition camera, extracts the lines in the main and side views of the abnormal positions, analyzes the parameters of the line curvature and distance in the main and side views, determines the types of the abnormal sites of the slope, including depression, protrusion, and crack, and determines the abnormalities of the slope in combination with the remaining same abnormal parameters under the same soil characteristics in the same group; The potential anomaly determination module marks the potential abnormal positions manually through the slope images collected by the high-definition camera, extracts the data of the potential abnormal positions and reconciles them with the lidar point cloud data, and monitors and identifies the potential abnormalities and similar abnormalities of the slope based on the data upload results of the potential abnormal positions.

Images