CN-121997156-A - Landslide vulnerability evaluation method based on support vector machine model

CN121997156ACN 121997156 ACN121997156 ACN 121997156ACN-121997156-A

Abstract

The invention discloses a landslide vulnerability assessment method based on a support vector machine model, which belongs to the technical field of geological disaster assessment and comprises the following steps of S1, collecting basic data of a region to be studied, processing and analyzing the basic data to construct an index system, S2, randomly selecting equal numbers of non-landslide samples in a non-landslide region based on distribution of known landslide points, S3, training a support vector machine based on the landslide samples to obtain an optimal support vector machine model, and S4, predicting the landslide vulnerability of the region to be studied by utilizing the optimal support vector machine model, and generating a landslide vulnerability partition map according to a prediction result. The landslide susceptibility evaluation method provided by the invention improves the accuracy of landslide susceptibility evaluation results, and can provide reference basis for geological disaster prevention and control work, so that the landslide susceptibility evaluation method is high in reliability, good in accuracy, objective and scientific.

Inventors

LI RUIDONG
ZHANG FANYU
WANG YUJIA
Wang Jinglang
HAN BINGBING
DING GUOXUAN

Assignees

兰州大学
甘肃工程地质研究院

Dates

Publication Date: 20260508
Application Date: 20260123

Claims (7)

1. A landslide vulnerability evaluation method based on a support vector machine model is characterized by comprising the following steps: S1, collecting basic data of a region to be researched, processing and analyzing the basic data, and constructing an index system; S2, randomly selecting equal numbers of non-landslide samples in a non-landslide area based on the distribution of known landslide points; S3, training a support vector machine based on sample points and an index system of non-landslide samples to obtain an optimal support vector machine model; S4, predicting landslide susceptibility of the area to be researched by utilizing the optimal support vector machine model, and generating a landslide susceptibility partition map according to a prediction result.
2. The landslide vulnerability assessment method based on support vector machine model of claim 1, wherein S1 comprises the substeps of: s11, collecting basic data of a region to be researched, and standardizing the basic data; s12, performing principal component analysis and Szelman correlation analysis on the standardized basic data to obtain a plurality of evaluation factors; s13, screening stratum lithology, landform type, earthquake peak acceleration, road density, fault density, normalized vegetation index, river network density, annual average precipitation, land type, slope direction and elevation from a plurality of evaluation factors, and constructing an index system.
3. The method for evaluating the susceptibility to landslide of the support vector machine model according to claim 1, wherein in S2, based on the known landslide sample points, a non-landslide sample equal to the landslide point is generated in a non-landslide region by using a random generation point function of ArcGIS.
4. The landslide vulnerability assessment method based on support vector machine model of claim 1, wherein S3 comprises the substeps of: S31, extracting attributes of a plurality of evaluation index factors of an index system to landslide sample points; s32, constructing a support vector machine, and determining an optimization problem of the support vector machine; s33, processing the optimization problem of the support vector machine to obtain the latest optimization problem; And S34, training the sample points by using a support vector machine algorithm based on the latest optimization problem to obtain an optimal support vector machine model.
5. The landslide vulnerability evaluation method based on support vector machine model of claim 4, wherein in S31, the expression of the support vector machine is: ; Wherein, the Representing the normal vector of the hyperplane, Showing the offset top of the container, Representing the feature vector of a sample.
6. The landslide vulnerability evaluation method based on support vector machine model of claim 4, wherein the expression of the optimization problem of the support vector machine is: 2 1, ; Wherein, the Representing the normal vector of the hyperplane, Showing the offset top of the container, Represent the first The class labels of the individual samples are used, Represent the first The feature vector of the individual samples is used, Representing an index of the samples in the dataset.
7. The landslide vulnerability assessment method based on support vector machine model of claim 4, wherein in S33, the expression of the latest optimization problem is: ; Wherein, the Representing the normal vector of the hyperplane, Showing the offset top of the container, The regularization parameters are represented by a set of values, Representing the total number of training samples, Represent the first The relaxation variables of the individual samples are, Represents a variable of the relaxation, Represent the first The class labels of the individual samples are used, An index representing a sample in the dataset, Represent the first Feature vectors of the individual samples.

Description

Landslide vulnerability evaluation method based on support vector machine model Technical Field The invention belongs to the technical field of geological disaster evaluation, and particularly relates to a landslide vulnerability evaluation method based on a support vector machine model. Background The topography of China is high and low in east and west, the landform types are various, and various geological disasters are frequent in sand storm, loess and karst landform distribution areas. The Chinese operators are wide, the relief is large, the territory area is large, the natural geological environment condition for inoculating the geological disasters is complex, and the nature and the intensity of human engineering activities in different areas are different, so that the formed geological disasters have large differences in type, development intensity and hazard size, and the Chinese has become one of countries with frequent geological disasters in the world. In our country, landslide disasters are widely distributed, so landslide disasters are the focus of our study on geological disasters. The susceptibility evaluation of geological disasters is to study the probability of geological disasters in a certain area, and is usually carried out by adopting extremely high, medium, low and extremely low susceptibility grades. If the landslide is accurately predicted before the landslide occurs, corresponding prevention measures can be timely taken, so that the possible loss and damage to people caused by landslide disasters are avoided or reduced to a certain extent. In recent years, with the development of remote sensing technology, geographic information systems and computer technology, the accuracy of landslide susceptibility evaluation results is also continuously improved. The information quantity model is applied to many actual production units, and the weight relation among all evaluation factors may not be fully considered by the method, which may lead to inaccuracy of the result of the vulnerability zoning, so that the machine learning method is being widely popularized and used. Therefore, the study selects a machine model for susceptibility evaluation. Disclosure of Invention In order to solve the problems, the invention provides a landslide vulnerability evaluation method based on a support vector machine model. The technical scheme of the invention is that the landslide vulnerability evaluation method based on the support vector machine model comprises the following steps: S1, collecting basic data of a region to be researched, processing and analyzing the basic data, and constructing an index system; S2, randomly selecting equal numbers of non-landslide samples in a non-landslide area based on the distribution of known landslide points; S3, training a support vector machine based on sample points and an index system of non-landslide samples to obtain an optimal support vector machine model; S4, predicting landslide susceptibility of the area to be researched by utilizing the optimal support vector machine model, and generating a landslide susceptibility partition map according to a prediction result. Further, S1 comprises the following sub-steps: s11, collecting basic data of a region to be researched, and standardizing the basic data; s12, performing principal component analysis and Szelman correlation analysis on the standardized basic data to obtain a plurality of evaluation factors; s13, screening stratum lithology, landform type, earthquake peak acceleration, road density, fault density, normalized vegetation index, river network density, annual average precipitation, land type, slope direction and elevation from a plurality of evaluation factors, and constructing an index system. Further, in S2, based on the known landslide sample points, a non-landslide sample equal to the landslide point is generated in the non-landslide region by using the random generation point function of ArcGIS. Further, S3 comprises the following sub-steps: S31, extracting attributes of a plurality of evaluation index factors of an index system to landslide sample points; s32, constructing a support vector machine, and determining an optimization problem of the support vector machine; s33, processing the optimization problem of the support vector machine to obtain the latest optimization problem; And S34, training the sample points by using a support vector machine algorithm based on the latest optimization problem to obtain an optimal support vector machine model. Further, in S31, the expression of the support vector machine is: ; Wherein, the Representing the normal vector of the hyperplane,Showing the offset top of the container,Representing the feature vector of a sample. Further, the expression of the optimization problem of the support vector machine is: 21,; Wherein, the Representing the normal vector of the hyperplane,Showing the offset top of the container,Represent the firstThe class labels of the indi