Post-fire slope optimization method

Abstract

The invention discloses a post-fire slope optimization method, which belongs to the field of slope rockfall blocking, and comprises the following steps: S1, constructing a multi-source data set; s2, determining a rockfall propagation range; s3, constructing a disaster-exposure-fragility three-dimensional risk assessment model, and outputting a slope risk grade map and a risk score; s4, based on a risk assessment result, a fused interaction matrix and a weight calculation method, introducing a risk-barrier energy dynamic adaptation model, constructing a site selection adaptability-protection energy-structure parameter dynamic optimization system, and outputting a flexible barrier site selection scheme and barrier parameters; and S5, based on a real-time monitoring result, evaluating a flexible barrier site selection scheme and barrier parameters, and dynamically optimizing the step S4 according to an evaluation result. By adopting the post-fire slope optimization method, through integration of multi-source monitoring data and a dynamic simulation technology, the limitation that a traditional method is insufficient in adaptation to special working conditions after the fire is broken through, and the dual goals of safety protection and ecological restoration are achieved.

Description

A method for optimizing slopes after a fire Technical Field

[0001] This invention relates to the field of slope rockfall prevention technology, and in particular to a method for optimizing slopes after a fire. Background Technology

[0002] Post-fire slopes are prone to rockfalls due to vegetation damage, changes in rock mechanics, and intensified deep deformation, posing a serious threat to surrounding infrastructure, the ecological environment, and human activities. With global warming, the frequency and intensity of extreme fire events have significantly increased, making post-fire slope disaster prevention and control a key requirement in the field of geological disaster management. However, existing technologies still have many limitations in monitoring, assessment, and protection, making it difficult to meet the "precise, quantitative, and sustainable" management requirements in engineering practice. Current monitoring methods for post-fire slopes are mainly divided into remote sensing monitoring and field investigation, but both technologies have significant shortcomings when applied alone. The limitations of macroscopic remote sensing technologies (InSAR, LiDAR) are as follows: Although InSAR technology can achieve long-term, large-scale monitoring of deep slope displacement (such as quantification of annual displacement rate), its spatial resolution is low (usually at the meter level), and it cannot accurately capture key parameters such as rock joint orientation and microscopic features of vegetation damage. Furthermore, it is not sensitive enough to the short-term vegetation recovery dynamics after a fire. LiDAR technology can generate high-precision digital terrain models (DTM), but it can only reflect the macroscopic morphology of the terrain and lacks the ability to identify parameters such as rock mechanical properties (such as compressive strength) and the volume distribution of falling rocks.

[0003] Limitations of microscopic field surveys and UAV technology: While field surveys can obtain microscopic data such as rock mechanics parameters and rockfall size, their coverage is limited. Furthermore, after a fire, slopes often have steep terrain and difficult access, resulting in low survey efficiency and insufficient data representativeness. Although UAV photogrammetry can supplement microscopic geological information, it cannot monitor deep slope displacement, making it difficult to reflect the dynamic changes in the overall stability of the slope after a fire. Moreover, the quality of data acquisition is easily affected by weather conditions (such as smoke and precipitation). Summary of the Invention

[0004] The purpose of this invention is to provide a method for optimizing slopes after a fire, thereby solving the aforementioned technical problems.

[0005] To achieve the above objectives, this invention provides a method for optimizing slopes after a fire, comprising the following steps: S1, constructing a multi-source dataset by integrating post-fire slope topography, geology, vegetation, and deep deformation, and combining it with field survey data; S2, quantifying initial rockfall parameters based on the multi-source dataset obtained in S1, conducting 3D trajectory simulation through vegetation-block interaction simulation combined with a rolling friction model, and quantifying the impact of vegetation damage on rockfall propagation using reachability angle theory to determine the rockfall propagation range; S3, based on the rockfall propagation range determined in step S2, introducing InS on the basis of the FiRo-RiS system. Using AR displacement factor and target object size effect, a three-dimensional risk assessment model of disaster-exposure-vulnerability is constructed, outputting slope risk level map and risk score; S4, based on the risk assessment results output in S3, the interaction matrix and weight calculation method are integrated, and a dynamic adaptation model of risk-barrier energy is introduced to construct a dynamic optimization system of site selection adaptability-protection energy-structural parameters, outputting flexible barrier site selection scheme and barrier parameters; S5, based on real-time monitoring results, the flexible barrier site selection scheme and barrier parameters are evaluated, and the flexible barrier site selection scheme and barrier parameters output in step S4 are dynamically optimized according to the evaluation results.

[0006] Therefore, the above-mentioned post-fire slope optimization method of this invention has the following beneficial effects: 1. Significantly improved risk assessment accuracy: By integrating key geological parameters after a fire, such as joint density and rock mass compressive strength, and combining three-dimensional rockfall trajectory simulation and dynamic risk scoring model, the potential rockfall source area, movement path, and high-risk area are accurately identified, avoiding omissions or misjudgments caused by traditional static assessments, and providing a scientific basis for protection design; 2. Enhanced targeting and effectiveness of protection engineering: Based on the quantitative results of post-fire rockfall kinetic energy and trajectory length, the core parameters such as the flexible barrier installation angle, protection energy, and anchor bolt length are optimized, making the protection structure perpendicular to the mainstream rockfall movement direction, increasing the interception success rate by 10%-15%, and effectively resisting the frequent high-kinetic-energy and dispersed rockfall disasters after a fire; 3. Achieving synergy between engineering protection and ecological restoration: The scheme takes into account both short-term disaster interception and long-term slope stability. The method utilizes fire debris (broken wood, branches, and ash) to construct a vegetation-soil-gravel composite structure, restoring the slope's erosion resistance and water retention capacity. This addresses the issue of instability caused by traditional protection methods that focus solely on engineering reinforcement while neglecting ecological restoration. Furthermore, it optimizes engineering economy and construction feasibility by precisely optimizing parameters to avoid over- or under-design, reducing material waste and post-construction maintenance costs. It is also adaptable to slopes with varying degrees of fire damage (mild to severe) and slope gradients (20°-60°), supporting local material sourcing and standardized construction, thus reducing construction difficulty and cost. Finally, it expands applicability and practicality by overcoming the limitations of traditional methods in adapting to specific geological conditions after fires (soil aggregate disintegration and vegetation root carbonization). This method can be widely applied to mountain roads, railway slopes, and forest fire-affected areas, providing a replicable and scalable technical paradigm for post-fire slope disaster prevention.

[0007] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0008] Figure 1 is a flowchart of a post-fire slope optimization method according to the present invention. Detailed Implementation

[0009] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the embodiments of the present invention and are not intended to limit the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of this application. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout.

[0010] It should be noted that the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, such as a process, method, system, product, or server that includes a series of steps or units, not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such process, method, product, or device.

[0011] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0012] As shown in Figure 1, a method for optimizing slopes after a fire includes the following steps: S1, constructing a multi-source dataset by integrating the topography, geology, vegetation, and deep deformation of the post-fire slope and combining it with field survey data; In step S1, the topographic slope data acquisition steps are as follows: First, collect point cloud data using 1m resolution airborne LiDAR, and then use statistical filtering to calculate the elevation difference between each point and its neighboring points: In the formula, Represents coordinate points The elevation difference between it and its neighboring points; Represents coordinate points Elevation; Represents coordinate points The average elevation of the neighborhood points; then, using rules Noise points exceeding the threshold are removed to obtain the filtered elevation. ; The elevation difference threshold is represented; then, a digital terrain model is generated using Kriging interpolation, and the slope of each grid cell is calculated. : In the formula, and They represent , Elevation gradient in direction; finally, screening. The area is considered a potential source of rockfall. This indicates the slope threshold.

[0013] The steps for acquiring geological, vegetation, and deep deformation data are as follows: First, 12 ground control points covering key geological boundaries are evenly distributed in the post-fire slope area. These key geological boundaries include the combustion residue accumulation area, the slope source area, the middle of the slope, and the toe. Then, the flight path of the UAV carrying the camera is planned: Step 1, delineation of the flight area boundary and key target points; importing the digital terrain model, potential rockfall source area vector map, and InSAR instability zoning map, overlaying them in GIS software, and extending the flight buffer margin outwards to generate the flight area boundary; the division rules for the InSAR instability zoning map are as follows: In the formula, The annual displacement rate of permanent scatterers on the slope after a fire is represented, and these permanent scatterers are extracted from InSAR data using coherent pixel technology. Key target points are marked, including potential rockfall source areas, unstable areas in InSAR, slope cliffs, areas with severe vegetation damage, and areas with block accumulation at the slope toe. Each target point has a bounding box with a diameter ≥2m to ensure priority coverage by the flight path. Restricted areas are excluded: areas with thick accumulations of combustion residue, temporary rockfall accumulation areas, and no-fly zones for drones are marked. The second step involves determining flight parameters, considering camera parameters and target accuracy requirements; these parameters include flight altitude. ,speed Overlap factor Flight intervals and the total number of images required : ; ; ; ; In the formula, Indicates the target ground sampling distance; Indicates the size of a single pixel in the camera sensor; Indicates the camera's focal length; Indicates the ground sampling distance, and ; Indicates the heading overlap (the percentage of overlap between two adjacent heading images); Indicates the camera shutter frequency; Indicates lateral overlap (the percentage of overlap between two adjacent flight path images); Indicates the ground coverage width of a single image; Indicates the total area of ​​the flight area; and These represent the horizontal and vertical pixel counts of the camera sensor, respectively. The third step involves designing flight path layouts and modes adapted to the post-fire slope conditions. Orthogonal flight path layout: For gently sloping areas, a grid-like flight path is used, parallel to the slope direction, covering all key target points. Oblique flight path layout: For steep cliffs and potential rockfall sources, three oblique flight paths with inclination angles of 30°, 45°, and 60° are set up to photograph the cliff face from different angles, avoiding obscuring joint features. The oblique and orthogonal flight paths overlap by ≥30% to ensure data fusion consistency. Flight path connection: A transition zone of ≥50m is set at the junction of orthogonal and oblique flight paths to avoid image stitching gaps. Key target points are covered by a circling flight path to ensure 360° coverage. The fourth step involves verifying coverage and accuracy: Overlaying flight paths and flight area boundaries, along with key target points, in GIS software to check for blind spots. If blind spots are found, supplementary flight paths are added. Ground sampling distance is also considered. (The actual ground distance corresponding to a single pixel in UAV imagery) and effective overlap factor Pre-evaluation of SfM point cloud resolution (Average spatial distance between two adjacent valid points in the point cloud): ;in, In the formula, This represents the point cloud density correction factor; Indicates the average tilt angle of the camera; Indicates the number of orthogonal flight paths; Indicates the camera tilt angle in orthographic mode; Indicates the number of oblique firing flight path groups; and They represent the first Number of oblique firing paths and camera tilt angle; This represents the total number of flight paths. Then, the UAV flies along the planned flight path, acquiring images using the Phase One iXM-100RGB camera during flight. After filtering out invalid images, these images are imported into Agisoft Metashape software for image preprocessing. The SIFT algorithm is used to extract key feature points from each image, and then the FLANN matching algorithm is used to match feature points across images, generating initial matching point pairs. Finally, using the exterior orientation elements, attitude angles, and 3D coordinates of the matching points from all images as optimization variables, the image projection error is minimized to obtain the optimized sparse point cloud. The optimization objective function is as follows: In the formula, This represents the sum of squares of the total projection errors; Indicates the number of valid images; Indicates the first The number of matching points on a valid image; Indicates the first Zhang Youyou's Image (First) The actual pixel coordinates of each matching point; Indicates the first Zhang Youyou's Image (First) The theoretical projected pixel coordinates of each matching point, and , , , These represent the image pose rotation matrices respectively. The first column; , , These represent the image pose rotation matrices respectively. The second column; , , These represent the image pose rotation matrices respectively. The third column; Indicates the coordinates of the camera's principal point; Indicates the first The three-dimensional coordinates of the matching points; Indicates the first The coordinates of the photographic center of Zhang Youzheng's image; among them... , , and These represent the pitch, yaw, and lateral angles of the images directly measured by the inertial measurement unit (IMU). Then, based on the optimized sparse point cloud, a multi-view stereo matching algorithm is used to estimate the pixel-level depth of each image, generating a depth map. The depth maps of all images are then fused, and the gaps in the point cloud are filled using a Poisson reconstruction algorithm to generate a dense point cloud with a resolution ≤5cm, preserving microscopic features such as rock joints and vegetation remnants. The dense point cloud is then converted to the WGS 84 UTM coordinate system, ensuring a one-to-one correspondence between the point cloud and the actual geographical location of the slope, resulting in a georegistered dense point cloud. Redundant noise points are then removed, and the final georegistered dense point cloud is output. Joint surfaces are extracted using CloudCompare to obtain the dip angle, dip direction, and number of each joint. Finally, using a 4m×4m sample method, the cross-sectional area density of surviving vegetation at breast height (BHH) within each sample plot is calculated, and the vegetation density is determined. and damage rate : ; In the formula, This represents the average cross-sectional area at breast height of the quadrat, and , This indicates the total number of quadrats. Indicates the first Total cross-sectional area of ​​chest height in each sample plot Indicates the first Number of surviving plants in each quadrat Indicates the first The topographic area at breast height of surviving vegetation. , Indicates the first Diameter at breast height (DBH) of surviving vegetation; Indicates the area of ​​a single sample plot; Indicates the vegetation coverage before the fire; This indicates the vegetation coverage after a fire.

[0014] S2. Based on the multi-source dataset obtained in S1, quantify the initial parameters of the falling rocks. Conduct 3D trajectory simulation through vegetation-block interaction simulation combined with a rolling friction model. Quantify the impact of vegetation damage on the propagation of falling rocks using the reachability angle theory to determine the propagation range of the falling rocks. Step S2 specifically includes the following steps: S21. Combine on-site statistics and remote sensing data to determine the initial parameters of the falling rocks, including the average block mass. and initial release rate : ; In the formula, Indicates the density of the rock mass; This represents the average block volume, which is obtained through on-site statistics. Represents gravitational acceleration; Indicates the average height of potential rockfall source areas; This represents the coefficient of friction, and , S22. 3D simulation of rockfall trajectory: RockyFor3D is used to simulate the energy dissipation of vegetation on the block, while the rolling friction model of stnParabel is used to correct the contact between the block and the ground. The intersection of the RockyFor3D and stnParabel trajectories is taken as the final effective rockfall trajectory; S23. Based on the effective rockfall trajectory, and considering vegetation density... Average block volume and slope of potential rockfall source areas Calculate the reach angle : ;in, ; ; In the formula, Indicates the basic reachability angle when there is no vegetation; Indicates the vegetation-corrected reachability angle; Indicates the slope of the potential rockfall source area; S24: Indicate the slope threshold of the potential rockfall source area; estimate the longest effective rockfall trajectory length. : S25, combined with the area of ​​potential rockfall source areas Calculate the frequency of rockfall release from potential rockfall source areas. Annual frequency of target objects being hit by falling rocks : ; In the formula, Indicates the reference annual release frequency; This represents the number of trajectories that reach the target object; This represents the total number of simulated trajectories.

[0015] S3. Based on the rockfall propagation range determined in step S2, and building upon the FiRo-RiS system, introduce the InSAR displacement factor and target object size effect to construct a three-dimensional risk assessment model of disaster-exposure-vulnerability, outputting a slope risk level map and risk score; step S3 specifically includes the following steps: S31. Calculate the disaster factor: ;in, ; ; ; ; In the formula, , , , , , The scores represent the slope height, block size, annual rainfall, rock mass quality, rockfall frequency, and displacement, respectively. Indicates annual rainfall; Table 1: Rock mass quality classification table

[0016] S32. Introduce the target object size effect to calculate the exposure-vulnerability factor. : ;in, ; In the formula, Indicates the exposure factor; This indicates the vulnerability factor (the potential for damage to a target object due to a rockfall). Indicates the importance level of the target object; , , , and These represent the five preset importance levels, from low to high; Indicates the distance from the toe of the slope to the target object; Indicates the width of the target object; S33, Calculates the risk score. And classify the risk levels: ; .

[0017] S4. Based on the risk assessment results output in S3, the interaction matrix and weight calculation method are integrated, and a dynamic adaptation model of risk-barrier energy is introduced to construct a dynamic optimization system of site selection adaptability-protection energy-structural parameters, outputting flexible barrier site selection schemes and barrier parameters; Step S4 specifically includes the following steps: S41. Construct an interaction matrix containing assessment indicators and calculate the site selection adaptability index. : ;in, In the formula, Indicates the first Weights of each evaluation indicator; Indicates the first Scores for each evaluation indicator , These respectively indicate unsuitable sites and suggested sites; Indicates the number of evaluation indicators; Indicates the first The values ​​for each evaluation indicator are obtained through expert scoring; Indicates the first The effectiveness values ​​of each evaluation indicator are obtained through expert scoring; S42, based on risk scoring and the maximum kinetic energy of falling rocks Dynamically calculate the protective energy required by the barrier. : ;in, In the formula, Indicates the first The kinetic energy of the trajectory of the falling rock, and , Indicates the first The resistance to the movement of falling rocks Indicates the first The longest trajectory length of a falling rock. Indicates the first The volume of the falling rock; S43, combined with the maximum bounce height of the falling rock. Design barrier height according to terrain slope With installation angle : ; In the formula, Indicates the slope of the installation location.

[0018] S5. Based on the real-time monitoring results, evaluate the flexible barrier site selection scheme and barrier parameters, and dynamically optimize the flexible barrier site selection scheme and barrier parameters output in step S4 according to the evaluation results.

[0019] Step S5 specifically includes the following steps: S51, Conduct UAV and InSAR monitoring every quarter, compare with the initial data, and calculate the vegetation restoration rate. and displacement change : ; In the formula, Indicates quarterly vegetation cover; Indicates the quarterly displacement rate; S52: Recalculate the risk score based on the updated monitoring data, representing the initial displacement rate. ,contrast S53. Obtain risk change trends and output risk change reports; based on the updated risk score Optimize flexible barrier site selection schemes and barrier parameters: ; ; ; In the formula, This represents the optimized site selection suitability index; and They represent the first Updated weights and scores for each evaluation indicator; This indicates the optimized height of the flexible barrier; Indicates joint density; This indicates the optimized installation angle; Indicates the correction term, and , and All represent conversion factors. It represents the uniaxial compressive strength of the rock mass.

[0020] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for optimizing slopes after a fire, characterized in that: Includes the following steps: S1. By integrating post-fire slope topography, geology, vegetation, and deep deformation, and combining them with field survey data, a multi-source dataset is constructed. S2. Based on the multi-source dataset obtained in S1, quantify the initial parameters of the falling rocks, conduct 3D trajectory simulation by combining vegetation-block interaction simulation with rolling friction model, and quantify the impact of vegetation damage on the propagation of falling rocks by combining reachability angle theory to determine the propagation range of falling rocks; S3. Based on the propagation range of falling rocks determined in step S2, on the basis of FiRo-RiS system, introduce InSAR displacement factor and target object size effect to construct a three-dimensional risk assessment model of disaster-exposure-vulnerability, and output slope risk level map and risk score; S4. Based on the risk assessment results output by S3, the interaction matrix and weight calculation method are integrated, and a risk-barrier energy dynamic adaptation model is introduced to construct a dynamic optimization system of site selection adaptability-protection energy-structural parameters, and output flexible barrier site selection scheme and barrier parameters; S5. Based on the real-time monitoring results, the flexible barrier site selection scheme and barrier parameters are evaluated, and the flexible barrier site selection scheme and barrier parameters output by step S4 are dynamically optimized according to the evaluation results.

2. The method for optimizing a slope after a fire according to claim 1, characterized in that: In step S1, the terrain slope data acquisition steps are as follows: First, 1m resolution airborne LiDAR point cloud data is collected, and then statistical filtering is used to calculate the elevation difference between each point and its neighboring points: In the formula, Represents coordinate points The elevation difference between it and its neighboring points; Represents coordinate points Elevation; Represents coordinate points The average elevation of the neighborhood points; then, using rules Noise points exceeding the threshold are removed to obtain the filtered elevation. ； The elevation difference threshold is represented; then, a digital terrain model is generated using Kriging interpolation, and the slope of each grid cell is calculated. ： In the formula, and They represent 、 Elevation gradient in direction; finally, screening. The area is considered a potential source of rockfall. This indicates the slope threshold.

3. The method for optimizing a slope after a fire according to claim 2, characterized in that: The steps for acquiring geological, vegetation, and deep deformation data are as follows: First, 12 ground control points covering key geological boundaries are evenly distributed in the post-fire slope area. These key geological boundaries include the combustion residue accumulation area, the slope source area, the middle of the slope, and the toe. Then, the flight path of the UAV carrying the camera is planned: Step 1, delineation of the flight area boundary and key target points; importing the digital terrain model, potential rockfall source area vector map, and InSAR instability zoning map, overlaying them in GIS software, and extending the flight buffer margin outwards to generate the flight area boundary; the division rules for the InSAR instability zoning map are as follows: In the formula, The annual displacement rate of permanent scatterers on the slope after a fire is represented, and these permanent scatterers are extracted from InSAR data using coherent pixel technology. Key target points are marked, including potential rockfall source areas, unstable areas in InSAR, slope cliffs, areas with severe vegetation damage, and areas with block accumulation at the slope toe. Each target point has a bounding box with a diameter ≥2m to ensure priority coverage by the flight path. Restricted areas are excluded: areas with thick accumulations of combustion residue, temporary rockfall accumulation areas, and no-fly zones for drones are marked. The second step involves determining flight parameters, considering camera parameters and target accuracy requirements; these parameters include flight altitude. ,speed Overlap factor Flight intervals and the total number of images required ： ； ； ； ； In the formula, Indicates the target ground sampling distance; Indicates the size of a single pixel in the camera sensor; Indicates the camera's focal length; Indicates the ground sampling distance, and ； Indicates the degree of overlap in headings; Indicates the camera shutter frequency; Indicates lateral overlap; Indicates the ground coverage width of a single image; Indicates the total area of ​​the flight area; and These represent the horizontal and vertical pixel counts of the camera sensor, respectively. The third step involves designing flight path layouts and modes adapted to the post-fire slope conditions. Orthogonal flight path layout: For gently sloping areas, a grid-like flight path is used, parallel to the slope direction, covering all key target points. Oblique flight path layout: For steep cliffs and potential rockfall sources, three oblique flight paths with inclination angles of 30°, 45°, and 60° are set, with an overlap of ≥30% between the oblique and orthogonal flight paths. Flight path connection: A transition zone of ≥50m is set at the junction of the orthogonal and oblique flight paths, and a circling shooting path is set for key target points. The fourth step involves verifying coverage and accuracy: Overlaying the flight paths, flight area boundaries, and key target points in GIS software to check for blind spots. If blind spots are found, supplementary flight paths are added. Ground sampling distance is also considered. and the effective factor of overlap Pre-evaluation of SfM point cloud resolution ： ;in, In the formula, This represents the point cloud density correction factor; Indicates the average tilt angle of the camera; Indicates the number of orthogonal flight paths; Indicates the camera tilt angle in orthographic mode; Indicates the number of oblique firing flight path groups; and They represent the first Number of oblique firing paths and camera tilt angle; This represents the total number of flight paths. Then, the UAV flies along the planned flight path, acquiring images using the Phase One iXM-100 RGB camera during flight. After filtering out invalid images, these images are imported into Agisoft Metashape software for image preprocessing. The SIFT algorithm is used to extract key feature points from each image, and then the FLANN matching algorithm is used to match feature points across images, generating initial matching point pairs. Finally, using the exterior orientation elements, attitude angles, and 3D coordinates of the matching points from all images as optimization variables, the image projection error is minimized to obtain the optimized sparse point cloud. The optimization objective function is as follows: In the formula, This represents the sum of squares of the total projection errors; Indicates the number of valid images; Indicates the first The number of matching points on a valid image; Indicates the first Zhang Youyou's Image (First) The actual pixel coordinates of each matching point; Indicates the first Zhang Youyou's Image (First) The theoretical projected pixel coordinates of each matching point, and ， 、 、 These represent the image pose rotation matrices respectively. The first column; 、 、 These represent the image pose rotation matrices respectively. The second column; 、 、 These represent the image pose rotation matrices respectively. The third column; Indicates the coordinates of the camera's principal point; Indicates the first The three-dimensional coordinates of the matching points; Indicates the first The coordinates of the photographic center of Zhang Youzheng's image; among them... ， 、 and These represent the pitch, yaw, and lateral angles of the images directly measured by the inertial measurement unit (IMU). Based on the optimized sparse point cloud, a multi-view stereo matching algorithm is used to estimate the pixel-level depth of each image, generating a depth map. The depth maps of all images are then fused, and the gaps in the point cloud are filled using a Poisson reconstruction algorithm to generate a dense point cloud with a resolution ≤5cm. The dense point cloud is then transformed to the WGS 84 UTM coordinate system, ensuring a one-to-one correspondence between the point cloud and the actual geographical location of the slope, resulting in a georegistered dense point cloud. Redundant noise points are then removed, and the final georegistered dense point cloud is output. Joint surfaces are extracted using CloudCompare to obtain the dip angle, dip direction, and number of each joint. Finally, using a 4m×4m sample method, the density of surviving vegetation at breast height (BHH) within each sample plot is calculated to determine the vegetation density. and damage rate ： ； In the formula, This represents the average cross-sectional area at breast height of the quadrat, and ， This indicates the total number of quadrats. Indicates the first Total cross-sectional area of ​​chest height in each sample plot Indicates the first Number of surviving plants in each quadrat Indicates the first The topographic area at breast height of surviving vegetation. ， Indicates the first Diameter at breast height (DBH) of surviving vegetation; Indicates the area of ​​a single sample plot; Indicates the vegetation coverage before the fire; This indicates the vegetation coverage after a fire.

4. The method for optimizing a slope after a fire according to claim 3, characterized in that: Step S2 specifically includes the following steps: S21, combining on-site statistics and remote sensing data to determine the initial parameters of the rockfall, including the average mass of the rockfall blocks. and initial release rate ： ； In the formula, Indicates the density of the rock mass; This represents the average block volume, which is obtained through on-site statistics. Represents gravitational acceleration; Indicates the average height of potential rockfall source areas; This represents the coefficient of friction, and ， S22. 3D simulation of rockfall trajectory: RockyFor3D is used to simulate the energy dissipation of vegetation on the block, while the rolling friction model of stnParabel is used to correct the contact between the block and the ground. The intersection of the RockyFor3D and stnParabel trajectories is taken as the final effective rockfall trajectory; S23. Based on the effective rockfall trajectory, and considering vegetation density... Average block volume and slope of potential rockfall source areas Calculate the reach angle ： ;in, ； ； In the formula, Indicates the basic reachability angle when there is no vegetation; Indicates the vegetation-corrected reachability angle; Indicates the slope of the potential rockfall source area; S24: Indicate the slope threshold of the potential rockfall source area; estimate the longest effective rockfall trajectory length. ： S25, combined with the area of ​​potential rockfall source areas Calculate the frequency of rockfall release from potential rockfall source areas. Annual frequency of target objects being hit by falling rocks ： ； In the formula, Indicates the reference annual release frequency; This represents the number of trajectories that reach the target object; This represents the total number of simulated trajectories.

5. A method for optimizing a slope after a fire, as described in claim 4, characterized in that: Step S3 specifically includes the following steps: S31, Calculate the disaster factor: ;in, ； ； ； ； In the formula, 、 、 、 、 、 The scores represent the slope height, block size, annual rainfall, rock mass quality, rockfall frequency, and displacement, respectively. S32 represents annual rainfall; Calculate the exposure-vulnerability factor by introducing the target object size effect. ： ;in, ； In the formula, Indicates the exposure factor; Indicates vulnerability factor; Indicates the importance level of the target object; 、 、 、 and These represent the five preset importance levels, from low to high; Indicates the distance from the toe of the slope to the target object; Indicates the width of the target object; S33, Calculates the risk score. And classify the risk levels: ； 。 6. A method for optimizing a slope after a fire, as described in claim 5, characterized in that: Step S4 Specifically, the steps include: S41, constructing an interaction matrix containing evaluation indicators and calculating the site selection adaptability index. ： ;in, In the formula, Indicates the first Weights of each evaluation indicator; Indicates the first Scores for each evaluation indicator ， These respectively indicate unsuitable sites and suggested sites; Indicates the number of evaluation indicators; Indicates the first The values ​​for each evaluation indicator are obtained through expert scoring; Indicates the first The effectiveness values ​​of each evaluation indicator are obtained through expert scoring; S42, based on risk scoring and the maximum kinetic energy of falling rocks Dynamically calculate the protective energy required by the barrier. ： ;in, In the formula, Indicates the first The kinetic energy of the trajectory of the falling rock, and ， Indicates the first The resistance to the movement of falling rocks Indicates the first The longest trajectory length of a falling rock. Indicates the first The volume of the falling rock; S43, combined with the maximum bounce height of the falling rock. Design barrier height according to terrain slope With installation angle ： ； In the formula, Indicates the slope of the installation location.

7. A method for optimizing a slope after a fire, as described in claim 6, characterized in that: Step S5 specifically includes the following steps: S51, Conduct UAV and InSAR monitoring every quarter, compare with the initial data, and calculate the vegetation restoration rate. and displacement change ： ； In the formula, Indicates quarterly vegetation cover; Indicates the quarterly displacement rate; S52: Recalculate the risk score based on the updated monitoring data, representing the initial displacement rate. ,contrast S53. Obtain risk change trends and output risk change reports; based on the updated risk score Optimize flexible barrier site selection schemes and barrier parameters: ； ； ； In the formula, This represents the optimized site selection suitability index; and They represent the first Updated weights and scores for each evaluation indicator; This indicates the optimized height of the flexible barrier; Indicates joint density; This indicates the optimized installation angle; Indicates the correction term, and ， and All represent conversion factors. It represents the uniaxial compressive strength of the rock mass.

Images