An engineering geological exploration and evaluation method based on remote sensing technology

Abstract

The present invention relates to a method for engineering geological exploration and evaluation based on remote sensing technology, belonging to the field of remote sensing technology. It uses lidar and unmanned aerial vehicle (UAV) aerial photography to obtain ground three-dimensional point cloud data, collects soil samples through geological exploration to obtain geological data, cleans and denoises the obtained point cloud data and geological data. According to the processed point cloud data, a digital elevation model is generated to analyze topographic features. Feature extraction is carried out on the point cloud data and geological data respectively, and a landslide risk assessment model is established using machine learning methods. According to the results of the model, the landslide risk of the target area is evaluated and graded, and the landslide risk assessment results are displayed. By combining point cloud data and geological data, the present invention can achieve the fusion of multi-source data, provide richer inputs for the landslide risk assessment model, and improve the accuracy of landslide risk assessment.

Description

Technical Field

[0001] The present invention relates to the field of remote sensing technology, and more specifically, to an engineering geological exploration and evaluation method based on remote sensing technology. Background Art

[0002] With the acceleration of the global urbanization process and the continuous expansion of infrastructure construction, engineering geological exploration plays an increasingly important role in the fields of civil engineering, environmental protection, resource development, etc. Traditional engineering geological exploration methods, such as on-site surveys and drilling, although they can provide relatively accurate geological data, their disadvantages are also very obvious: on the one hand, these methods require a long time cycle and high economic investment, and on the other hand, traditional methods mainly rely on point-like or limited-area survey data, making it difficult to comprehensively and systematically reflect the geological characteristics of a large area. Especially in areas with complex geological conditions and irregular terrain, the exploration difficulty is greater.

[0003] Remote sensing technology, as an efficient, fast, and economical geological exploration means, has received wide attention because it can cover large-scale surface information collection. In complex geological environments, remote sensing technology has higher acquisition efficiency and data processing capabilities, and can cover various geological problems such as groundwater, soil pollution, landslides, and ground fissures. Summary of the Invention

[0004] In view of the technical problems existing in the prior art, the present invention provides an engineering geological exploration and evaluation method based on remote sensing technology to solve the problems mentioned in the above background art.

[0005] The technical solution for solving the above technical problems of the present invention is as follows: An engineering geological exploration and evaluation method based on remote sensing technology specifically includes the following steps:

[0006] Step 101: Obtain ground three-dimensional point cloud data using lidar and UAV aerial photography, collect soil samples through geological exploration to obtain geological data, and clean and denoise the obtained point cloud data and geological data;

[0007] Step 102: Generate a digital elevation model based on the processed point cloud data to analyze terrain features;

[0008] Step 103: Extract features from the point cloud data and geological data respectively, and use machine learning methods to establish a landslide risk assessment model;

[0009] Step 104: Evaluate and classify the landslide risk of the target area according to the results of the model, and display the landslide risk assessment results.

[0010] In a preferred embodiment, in step 101, three-dimensional ground point cloud data is obtained using lidar and drone aerial photography. Geological data is acquired by conducting geological surveys to collect soil samples, and the obtained point cloud data and geological data are cleaned and denoised. The specific steps are as follows:

[0011] Step A1, data acquisition: The lidar scans the ground through laser pulses. According to the time difference of the returned signals, the three-dimensional coordinates of each point are measured, and the point cloud data is (x i , y i , z i ), representing the three-dimensional coordinates of the i-th point. The coordinates of each point include ground height information. A drone is used to carry a camera for aerial photography, recording position and attitude information, generating images and ground control points, and soil samples are collected through geological surveys, recording soil type, water content, and density to obtain geological data;

[0012] Step A2, data processing: The lidar point cloud is spatially aligned with the drone aerial photography image. Outlier processing is performed on the geological data and converted into a unified data format. By removing isolated points that do not belong to the ground, the quality of the point cloud is improved. It further includes the following steps:

[0013] Step A201, data registration: The iterative closest point algorithm is used to optimize the matching between the lidar point cloud and the drone image. Let P represent the set of points in the lidar point cloud, and Q be the set of points in the drone image. The goal is to minimize the error function By minimizing the distance between the lidar point cloud P and the drone image point cloud Q under the rotation matrix and displacement vector, the two are spatially aligned. Here, R is the rotation matrix, τ is the translation vector, P i is the i-th point in the lidar point cloud, Q i is the corresponding point in the drone image, and ||·|| represents the Euclidean norm of the vector;

[0014] Step A202, denoising: For the registered point cloud data, a ground extraction algorithm is used to remove non-ground points. A height threshold is set as h thershold , when the height of the point is greater than the threshold h thershold , it is determined that the point is a non-ground point, and all points above this threshold are deleted to ensure that the remaining point cloud only contains ground points.

[0015] In a preferred embodiment, in step 102, according to the processed point cloud data, a digital elevation model is generated to analyze terrain features. The specific steps are as follows:

[0016] Step B1, generating a digital elevation model: For the extracted ground points P j (x j, y j , z j ), whose height is z j is used as the height value in the digital surface model, and the point cloud data is converted into a regular grid composed of grid points (x g , y g ). For each grid point, from the extracted ground points P j (x j , y j , z j ), calculate the elevation as And repeat the interpolation calculation until the elevation values of all grid points have been calculated to generate a complete digital elevation model. Among them, w j is the weight calculated based on the distance between the ground point j and the grid point (x g , y g ), z g is the height value corresponding to the grid point (x g , y g ), representing the elevation of the digital elevation model, z j is the height value of the extracted ground points, and m is the number of ground points participating in the interpolation;

[0017] Step B2. Analyze the terrain features: Based on the generated digital elevation model, identify the landslide-prone areas by calculating the slope of each point in the digital elevation model. The formula for the slope at the grid point (x g , y g ) is By calculating the slope of each point in the entire area, obtain its slope value as S. Among them, S g is the slope of the grid point (x g , y g ), z is the elevation value of each point in the digital elevation model, is the elevation change rate of the point (x g , y g ) in the X direction in the digital elevation model, is the elevation change rate of the point (x g , y g ) in the Y direction in the digital elevation model.

[0018] In a preferred embodiment, in step 103, feature extraction is respectively performed on the point cloud data and the geological data, and a landslide risk assessment model is established by using a machine learning method. The specific steps are as follows:

[0019] Step C1. Feature extraction: Extract terrain features and geological features from the point cloud data and the geological data respectively, and combine the terrain features and the geological features to form a feature matrix Among them, each row represents a sample, and each column represents a feature. They respectively represent the slope value, soil type, water content of the sample, and soil density of the sample. N is the total number of samples, and the method further includes the following steps:

[0020] Step C101, extract terrain features: Extract the slope values of all grid points from the point cloud data and the generated digital elevation model to form a terrain feature array [S1, S2,..., S N ;

[0021] Step C102, extract geological features: Extract the soil type, water content, and density features from the geological data, expressed as Among them, is the soil type of the sample, W k is the water content of the sample, D k is the soil density of the sample, and k is the sample index;

[0022] Step C2, construct a landslide risk assessment model: Use the random forest algorithm for landslide risk assessment. Train the random forest model through the feature matrix X and the corresponding landslide risk label y. Among them, y = 1 indicates that a landslide occurs, and y = 0 indicates that no landslide occurs. The random forest constructs decision trees by calculating the splitting criteria of each tree. The specific calculation formula is: Among them, Gini(t) is the Gini impurity of node t, R is the number of categories, and p l is the proportion of category l in this node;

[0023] By constructing multiple decision trees, vote on each test sample, and finally predict the label of the sample. After the model training is completed, use the trained random forest model to predict new samples.

[0024] In a preferred embodiment, in step 104, the landslide risk of the target area is evaluated and graded according to the results of the model, and the landslide risk assessment results are displayed. The specific steps are as follows:

[0025] Step D1, landslide risk assessment: Evaluate the landslide risk of each sample in the target area, and use T decision trees for prediction. For each sample, calculate the probability that each tree predicts "landslide occurs" as Among them, P(slip|X test ) represents the landslide occurrence probability of sample X test in the trained model, λ t is the prediction output of the t-th tree, and T is the number of decision trees;

[0026] Step D2, risk level classification: Set the classification threshold of the risk level as P high = 0.5 and Plow = 0.3, traverse all samples in the target area and assign them to high-risk, medium-risk, and low-risk levels according to the calculated probabilities. When P(slip) > 0.5, the sample points are classified into the high-risk area; when 0.3 < P(slip) ≤ 0.5, the sample points are classified into the medium-risk area; when P(slip) ≤ 0.3, the sample points are classified into the low-risk area;

[0027] Step D3: Generate a risk map: Use GIS software to color-label each sample point according to its risk level, mark the high-risk area as red, the medium-risk area as yellow, and the low-risk area as green.

[0028] The beneficial effects of the present invention are as follows: Use lidar and drone aerial photography to obtain ground three-dimensional point cloud data, collect soil samples through geological exploration to obtain geological data, clean and denoise the obtained point cloud data and geological data, generate a digital elevation model based on the processed point cloud data to analyze terrain features, extract features from the point cloud data and geological data respectively to improve the accuracy of the evaluation model, and use machine learning methods to establish a landslide risk assessment model. Evaluate and classify the landslide risk of the target area according to the results of the model, and display the landslide risk assessment results. The present invention can quickly obtain geological data of a large area through remote sensing technology, reduce the dependence on on-site work in traditional exploration, and can achieve accurate landslide risk assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a flowchart of the method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0030] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts fall within the scope of protection of the present application.

[0031] In the description of the present application, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present application, "a plurality" means two or more unless otherwise specifically defined.

[0032] In the description of the present application, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described in the present application as "for example" is not necessarily to be construed as more preferred or advantageous than other embodiments. The following description is given to enable any person skilled in the art to implement and use the present invention. In the following description, details are set forth for purposes of explanation. It should be understood that those of ordinary skill in the art can recognize that the present invention can be implemented without the use of these specific details. In other instances, well-known structures and processes are not elaborated in detail so as not to obscure the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed in the present application.

[0033] Embodiment 1

[0034] This embodiment provides a method for engineering geological exploration and assessment based on remote sensing technology as shown in Figure 1 and specifically includes the following steps:

[0035] Step 101: Use lidar and UAV aerial photography to obtain ground three-dimensional point cloud data, collect soil samples through geological exploration to obtain geological data, and clean and denoise the obtained point cloud data and geological data;

[0036] Step 102: Generate a digital elevation model based on the processed point cloud data to analyze topographic features;

[0037] Step 103: Extract features from the point cloud data and geological data respectively, and use machine learning methods to establish a landslide risk assessment model;

[0038] Step 104: Evaluate and classify the landslide risk of the target area according to the results of the model, and display the landslide risk assessment results.

[0039] Preferably, in step 101, using lidar and UAV aerial photography to obtain ground three-dimensional point cloud data, collecting soil samples through geological exploration to obtain geological data, and cleaning and denoising the obtained point cloud data and geological data, the specific steps are as follows:

[0040] Step A1: Data acquisition: The lidar scans the ground through laser pulses, measures the three-dimensional coordinates of each point according to the time difference of the returned signals, and obtains the point cloud data as (x i , y i , z i), representing the three-dimensional coordinates of the \(i\)-th point, where the coordinates of each point include ground height information. An unmanned aerial vehicle (UAV) equipped with a camera is used for aerial photography to record position and attitude information, generate images and ground control points, and soil samples are collected through geological exploration to record soil type, water content, and density in order to obtain geological data;

[0041] Step A2, Data Processing: Align the lidar point cloud with the UAV aerial photography image in space, perform outlier processing on the geological data, and convert it into a unified data format. By removing isolated points that do not belong to the ground, the quality of the point cloud is improved. It further includes the following steps:

[0042] Step A201, Data Registration: Use the iterative closest point (ICP) algorithm to optimize the matching between the lidar point cloud and the UAV image. Let \(P\) represent the set of points in the lidar point cloud and \(Q\) be the set of points in the UAV image. The goal is to minimize the error function By minimizing the distance between the lidar point cloud \(P\) and the UAV image point cloud \(Q\) under the rotation matrix and displacement vector, the two are aligned in space. Here, \(R\) is the rotation matrix used to describe the rotation of the point cloud, and \(\tau\) is the translation vector used to describe the translation of the point cloud. \(P\) i is the \(i\)-th point in the lidar point cloud, and \(Q\) i is the corresponding point in the UAV image, and \(\|\cdot\|\) represents the Euclidean norm of the vector;

[0043] Step A202, Denoising: For the registered point cloud data, use a ground extraction algorithm to remove non-ground points. Set a height threshold of \(h\) thershold . When the height of a point is greater than the threshold \(h\) thershold , determine that the point is a non-ground point and delete all points above this threshold to ensure that the remaining point cloud only contains ground points.

[0044] Preferably, in step 102, according to the processed point cloud data, a digital elevation model (DEM) is generated to analyze terrain features. The specific steps are as follows:

[0045] Step B1, Generate Digital Elevation Model: For the extracted ground points \(P\) j (x j ,y j ,z j ), its height \(z\) j is used as the height value in the digital surface model. The point cloud data is converted into a regular grid composed of grid points \((x\) g ,y g ) through an interpolation method. For each grid point, from the extracted ground points \(P\) j (x j ,y j ,z j ) the elevation is calculated as And repeat the interpolation calculation until the elevation values of all grid points have been calculated to generate a complete digital elevation model, where w j is the weight calculated based on the distance between the ground point j and the grid point (x g , y g ), z g is the height value corresponding to the grid point (x g , y g ), representing the elevation of the digital elevation model, z j is the height value of the extracted ground points, and m is the number of ground points participating in the interpolation;

[0046] Step B2. Analyze the terrain features: Based on the generated digital elevation model, identify the landslide-prone areas by calculating the slope of each point in the digital elevation model. The formula for the slope at the grid point (x g , y g ) is By calculating the slope of each point in the entire area, obtain its slope value S, where S g is the slope of the grid point (x g , y g ), z is the elevation value of each point in the digital elevation model, is the elevation change rate of the point (x g , y g ) in the digital elevation model in the X direction, is the elevation change rate of the point (x g , y g ) in the digital elevation model in the Y direction.

[0047] Preferably, in step 103, feature extraction is respectively performed on the point cloud data and the geological data, and a landslide risk assessment model is established by using a machine learning method. The specific steps are as follows:

[0048] Step C1. Feature extraction: Extract the terrain features and geological features from the point cloud data and the geological data respectively, and combine the terrain features and the geological features to form a feature matrix where each row represents a sample and each column represents a feature, respectively represent the slope value, soil type, water content of the sample, soil density of the sample, and N is the total number of samples. Further, the following steps are included:

[0049] Step C101. Extract terrain features: Extract the slope values of all grid points from the point cloud data and the generated digital elevation model to form a terrain feature array [S1, S2,..., S N ;

[0050] Step C102, Extract Geological Features: Extract soil type, water content, and density features from geological data, expressed as where is the soil type of the sample, W k is the water content of the sample, D k is the soil density of the sample, and k is the sample index;

[0051] Step C2, Construct a Landslide Risk Assessment Model: Use the random forest algorithm for landslide risk assessment. Train the random forest model through the feature matrix X and the corresponding landslide risk label y. Among them, y = 1 indicates that a landslide has occurred, and y = 0 indicates that no landslide has occurred. The random forest constructs decision trees by calculating the splitting criteria of each tree. The specific calculation formula is: where Gini(t) is the Gini impurity of node t, R is the number of categories, and p l is the proportion of category l in this node;

[0052] By constructing multiple decision trees, vote on each test sample, and finally predict the label of the sample. After the model training is completed, use the trained random forest model to predict new samples.

[0053] Preferably, in step 104, evaluate and classify the landslide risk of the target area according to the results of the model, and display the landslide risk assessment results. The specific steps are as follows:

[0054] Step D1, Landslide Risk Assessment: Evaluate the landslide risk of each sample in the target area, and use T decision trees for prediction. For each sample, calculate the probability that each tree predicts "landslide occurs" as where P(slip|X test ) represents the landslide occurrence probability of sample X test in the trained model, λ t is the prediction output of the t-th tree, and T is the number of decision trees;

[0055] Step D2, Risk Level Classification: Set the classification thresholds for risk levels as P high = 0.5 and P low = 0.3. Traverse all samples in the target area, and assign them to high-risk, medium-risk, and low-risk levels according to the calculated probabilities. When P(slip) > 0.5, divide the sample point into the high-risk area. When 0.3 < P(slip) ≤ 0.5, divide the sample point into the medium-risk area. When P(slip) ≤ 0.3, divide the sample point into the low-risk area;

[0056] Step D3: Generate a risk map: Use GIS software to color-code each sample point according to its risk level, marking high-risk areas in red, medium-risk areas in yellow, and low-risk areas in green.

[0057] It should be noted that in the above embodiments, the descriptions of the various embodiments have their own focuses. For parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0058] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0059] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one or more of the processes or multiple processes and / or blocks Figure 1 one or more of the blocks or multiple blocks.

[0060] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in Figure 1 one or more of the processes or multiple processes and / or blocks Figure 1 one or more of the blocks or multiple blocks.

[0061] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are performed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one or more of the processes or multiple processes and / or blocks Figure 1 one or more of the blocks or multiple blocks.

[0062] Although the preferred embodiments of the present invention have been described, additional changes and modifications can be made to these embodiments by those skilled in the art once they learn of the basic inventive concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications that fall within the scope of the present invention.

[0063] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. An engineering geological exploration and evaluation method based on remote sensing technology, characterized in that, Specifically, it includes the following steps: Step 101: Obtain ground three-dimensional point cloud data using lidar and UAV aerial photography, collect soil samples through geological exploration to obtain geological data, and clean and denoise the obtained point cloud data and geological data; Step 102: Generate a digital elevation model based on the processed point cloud data to analyze terrain features. Among them, Step B1: Generate a digital elevation model: For the extracted ground points , their height is used as the height value in the digital surface model. The point cloud data is converted into a regular grid composed of grid points through an interpolation method. For each grid point, the elevation is calculated as from the extracted ground points , and the interpolation calculation is repeated until the elevation values of all grid points have been calculated to generate a complete digital elevation model. Among them, is the weight calculated based on the distance between the ground point j and the grid point , is the height value corresponding to the grid point , representing the elevation of the digital elevation model, is the height value of the extracted ground points, and m is the number of ground points participating in the interpolation; Step B2: Analyze terrain features: Based on the generated digital elevation model, identify landslide-prone areas by calculating the slope of each point in the digital elevation model. The formula for the slope at the grid point is . By calculating the slope of each point in the entire area, its slope value is obtained as S. Among them, is the slope of the grid point , z is the elevation value of each point in the digital elevation model, is the elevation change rate of the point in the X direction in the digital elevation model, is the elevation change rate of the point in the Y direction in the digital elevation model; Step 103: Extract features from the point cloud data and geological data respectively, and use machine learning methods to establish a landslide risk assessment model. Among them, Step C1: Feature extraction: Extract topographic features and geological features from the point cloud data and geological data respectively, and combine the topographic features and geological features to form a feature matrix , where each row represents a sample and each column represents a feature, representing slope value, soil type, water content of the sample, and soil density of the sample respectively, and N is the total number of samples; Step C2: Construct a landslide risk assessment model: Use the random forest algorithm for landslide risk assessment, and train the random forest model through the feature matrix X and the corresponding landslide risk label y. Among them, indicating that a landslide has occurred, indicating that it has not occurred. The random forest constructs decision trees by calculating the splitting criteria of each tree. The specific calculation formula is: , where is the Gini impurity of node t, R is the number of categories, is the proportion of category l in this node; By constructing multiple decision trees, vote on each test sample, and finally predict the label of the sample. After the model training is completed, use the trained random forest model to predict new samples; Step 104: Evaluate and classify the landslide risk of the target area according to the results of the model, and display the landslide risk assessment results.

2. The engineering geological exploration and evaluation method based on remote sensing technology according to claim 1, characterized in that: In the said Step 101, to obtain ground three-dimensional point cloud data using lidar and UAV aerial photography, collect soil samples through geological exploration to obtain geological data, and clean and denoise the obtained point cloud data and geological data, the specific steps are as follows: Step A1, data acquisition: The lidar scans the ground with laser pulses and measures the three-dimensional coordinates of each point based on the time difference of the returned signals, obtaining the point cloud data as , representing the three-dimensional coordinates of the i-th point. The coordinates of each point include ground height information. A drone is used to carry a camera for aerial photography, recording the position and attitude information, generating images and ground control points, and soil samples are collected through geological exploration, recording the soil type, water content, and density to obtain geological data; Step A2: Data processing: Align the lidar point cloud and the UAV aerial photography image spatially, process the outliers of the geological data, and convert them into a unified data format. By removing the isolated points that do not belong to the ground, the quality of the point cloud is improved.

3. A method for engineering geological exploration and evaluation based on remote sensing technology according to claim 2, characterized in that In the data processing of the said Step A2, to align the lidar point cloud and the UAV aerial photography image spatially, process the outliers of the geological data, and convert them into a unified data format. By removing the isolated points that do not belong to the ground, the quality of the point cloud is improved, which further includes the following steps: Step A201, Data Registration: Use the Iterative Closest Point (ICP) algorithm to optimize the matching between the lidar point cloud and the UAV image. Denote the set of points in the lidar point cloud as P and the set of points in the UAV image as Q. The goal is to minimize the error function , by minimizing the distance between the lidar point cloud P and the UAV image point cloud Q under the rotation matrix and displacement vector, so that the two are spatially aligned. Here, R is the rotation matrix, is the translation vector, is the i-th point in the lidar point cloud, is the corresponding point in the UAV image, represents the Euclidean norm of the vector; Step A202, Denoising: For the registered point cloud data, use a ground extraction algorithm to remove non-ground points. Set a height threshold of , when the height of a point is greater than the threshold , determine that the point is a non-ground point and delete all points higher than this threshold to ensure that the remaining point cloud only contains ground points.

4. A method for engineering geological exploration and assessment based on remote sensing technology according to claim 1, characterized in that, In the said Step C1 feature extraction, extract topographic features and geological features from the point cloud data and geological data respectively, which further includes the following steps: Step C101, extracting topographic features: Extract the slope values of all grid points from the point cloud data and the generated digital elevation model to form a topographic feature array ; Step C102, Extract Geological Features: Extract soil type, water content, and density features from geological data, expressed as , where is the soil type of the sample, is the water content of the sample, is the soil density of the sample, and k is the sample index.

5. A method for engineering geological exploration and evaluation based on remote sensing technology according to claim 1, characterized in that, In the said Step 104, to evaluate and classify the landslide risk of the target area according to the results of the model, and display the landslide risk assessment results, the specific steps are as follows: Step D1, Landslide Risk Assessment: Conduct landslide risk assessment on each sample in the target area, and use T decision trees for prediction. For each sample, calculate the probability that each tree predicts "landslide occurrence" as , where represents the landslide occurrence probability of sample in the trained model, is the prediction output of the t-th tree, and T is the number of decision trees; Step D2, Risk level classification: Set the classification threshold of the risk level to and . Traverse all samples in the target area and assign them to high-risk, medium-risk, and low-risk levels according to the calculated probability. When , divide the sample points into high-risk areas. When , divide the sample points into medium-risk areas. When , divide the sample points into low-risk areas; Step D3: Generate a risk map: Use GIS software to color-code each sample point according to its risk level, mark the high-risk area as red, the medium-risk area as yellow, and the low-risk area as green.

Images