Multi-source data fusion determination method for small watershed geological disaster monitoring distribution target area

Abstract

The invention discloses a multi-source data fusion determination method for a small watershed geological disaster monitoring distribution target area, and relates to the technical field of geological disaster monitoring. The method comprises the following steps: acquiring historical geological disaster data and geographic space data in a small watershed area research area, and establishing a database; constructing a refined evaluation unit for the research area; carrying out risk evaluation on the research area by adopting a machine learning algorithm; carrying out modeling on the research area by adopting numerical simulation and analyzing dangerousness; acquiring a multi-temporal SAR image of the research area by using a satellite; setting a deformation rate threshold value, and screening high-danger areas; a plurality of similar evaluation methods are integrated to analyze the risk of geological disasters in a small watershed area, and a high-risk area is determined as a preferred monitoring point distribution target area. According to the method provided by the invention, the accuracy, rationality and scientificity of small watershed area geological disaster monitoring distribution target region demarcation are effectively improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of geological disasters, and in particular relates to a method for determining a target area for geological disaster monitoring in a small watershed based on multi-source data fusion. Background Art

[0002] Geological hazard monitoring deployment is a core component of the disaster prevention and mitigation system. Its core goal is to scientifically identify high-risk areas for potential disasters and rationally allocate monitoring resources. Therefore, developing a scientific, systematic, and efficient method for comprehensively determining target areas for geological hazard monitoring deployment is crucial for the prediction and prevention of geological hazards.

[0003] While traditional geohazard monitoring and location methods have accumulated considerable experience through long-term practice, their limitations and shortcomings have gradually become apparent. These limitations are primarily manifested in the following aspects: Reliance on subjective expert judgment leads to significant discrepancies in the hazard risk classification of the same region, often resulting in a chaotic "one person, one plan" situation. Over-reliance on historical similarities makes it difficult to respond to new hazards or atypical geological conditions. Previous geohazard monitoring and location methods often relied on qualitative descriptions and analyses based on "high, medium, and low" risk levels, lacking the support of quantitative indicators such as probability and thresholds. Excessive focus on historical disaster sites or overt risk areas leads to neglect of potential medium- and low-risk areas.

[0004] In order to solve the above problems, the present invention proposes a method for determining target areas for small watershed geological disaster monitoring based on multi-source data fusion. Summary of the Invention

[0005] The purpose of the present invention is to propose a multi-source data fusion determination method for the target area of ​​geological disaster monitoring in a small watershed to solve the problems of the existing traditional methods of geological disaster monitoring point distribution methods proposed in the above background technology, which are mainly based on experience and lack quantitative indicators and quantitative support.

[0006] To achieve the above objectives, the present invention adopts the following technical solutions:

[0007] A method for determining a target area for geological disaster monitoring in a small watershed by fusing multi-source data comprises the following steps:

[0008] S1. Obtain historical geological disaster data and geospatial data of the study area and establish a database;

[0009] S2. Construct refined evaluation units (grid units or slope units) for the small watershed study area, extract the geospatial data obtained in S1 to the corresponding grid units, and establish a dataset for quantitative evaluation of the study area;

[0010] S3. Use machine learning algorithms to evaluate the risk of the study area and obtain the distribution of high-risk areas in the study area;

[0011] S4. Obtain geotechnical parameters of the study area;

[0012] S5. Use numerical simulation software to calculate the stability coefficients of all evaluation units in the study area, perform spatial analysis and visualization based on the calculation results, and obtain the distribution of high-risk areas in the study area;

[0013] S6. Use the SAR data platform to obtain multi-temporal SAR images of the study area;

[0014] S7. Set a deformation rate threshold for the study area, and define the evaluation units with deformation greater than the threshold as high-risk areas to obtain the high-risk distribution of the study area;

[0015] S8. Use multiple disaster point hazard assessment methods to overlay and analyze the hazard assessment results of small watershed area assessment units, and identify high-risk areas as the preferred monitoring target areas.

[0016] Preferably, the historical geological disaster data in S1 include disaster point data or surface data; the geographic space data include topographic factors, geological structure factors, hydrological factors, vegetation factors, meteorological factors, and human activity factors.

[0017] Furthermore, the topographic factors include but are not limited to elevation, slope, aspect, slope shape, slope position, slope variability, aspect variability, micro-relief, profile curvature, plane curvature, land use type, slope type, terrain relief, terrain roughness, surface incision depth, elevation variation coefficient, and terrain moisture index;

[0018] The geological structural factors include but are not limited to the distance from the fault and the age of the strata;

[0019] The hydrological factors include but are not limited to groundwater type, water flow dynamic index, sediment transport index, runoff modulus, distance from the river, and river network density;

[0020] The vegetation factor includes but is not limited to the normalized vegetation index;

[0021] The meteorological factors include but are not limited to multi-year average rainfall, 24-hour rainfall, and previous effective rainfall;

[0022] The human activity factors include but are not limited to POI kernel density, distance from roads, and distance from buildings.

[0023] Preferably, the regional refined grid unit in S2 is constructed by ARCGIS software, which extracts various factors in the geographic spatial data into evaluation units as a risk assessment factor dataset for the study area;

[0024] The risk assessment factor dataset is used as a dataset for quantitative evaluation of the study area.

[0025] Preferably, the S3 specifically includes the following contents:

[0026] S301. Construct a risk assessment model using a machine learning algorithm; use disaster site data as positive samples, randomly select an equal amount of non-disaster site data as negative samples, or select an optimized negative sample strategy to construct a training data set for the risk assessment model for model training;

[0027] S302, performing factor screening based on a factor screening method to eliminate redundant factors and optimizing hyperparameters of the model;

[0028] S303. Based on the risk assessment factor data set, an optimized risk assessment model is used to obtain risk mapping and obtain the distribution of high risk assessment units in the study area.

[0029] Furthermore, the machine learning algorithm in S301 includes but is not limited to a tree-based model, a support vector machine, an artificial neural network, and a gradient boosting framework, and can be specifically implemented in the form of a decision tree, a random forest, a support vector machine (SVM), an artificial neural network (ANN), XGBoost, or LightGBM, and the risk assessment model is constructed using one or a combination of two or more of the machine learning algorithms;

[0030] The factor screening methods in S302 include but are not limited to causal analysis, correlation analysis, principal component analysis, recursive feature elimination, and geographic factor detector, and the factor screening adopts one or a combination of two or more of the factor screening methods;

[0031] The optimization methods include but are not limited to Bayesian optimization, gradient descent algorithm, grid search, random search, and genetic algorithm, and the optimization adopts one or a combination of two or more of the optimization methods.

[0032] Preferably, the geotechnical parameters of the study area in S4 include but are not limited to internal friction angle, cohesion, density, pore water pressure coefficient, and gravity.

[0033] Furthermore, the methods for obtaining geotechnical parameters in the study area include but are not limited to indoor experiments: direct shear test, triaxial test, and ring shear test; and outdoor experiments: standard penetration test (SPT) and static penetration test (CPT).

[0034] Preferably, the S5 specifically includes the following contents:

[0035] S501. Modeling the study area using numerical simulation software to calculate the stability coefficients of all evaluation units within the study area; further, the numerical simulation software includes but is not limited to using Scoops3D for numerical simulation;

[0036] S502. Perform spatial analysis and visualization on the stability coefficient results of the evaluation unit to obtain the high-risk distribution in the study area.

[0037] Preferably, the SAR data platform in S6 includes but is not limited to Sentinel-1, ALOS-2 / PALSAR-2, and TerraSAR-X satellites.

[0038] Preferably, the S7 specifically includes the following contents:

[0039] S701. Use software to perform image registration and interferogram generation, extract PS points in StaMPS, estimate deformation rate and time series, and set deformation rate thresholds for the study area. Furthermore, the image registration and interferogram generation software includes but is not limited to Snap and GMTSAR. S702. Meteorological data includes but is not limited to ERA5.

[0040] S702. Use meteorological data to correct the atmospheric phase, output the deformation map and superimpose it on the evaluation unit, extract the units whose deformation rate exceeds the threshold to obtain the high-risk distribution of the study area.

[0041] Preferably, the S8 specifically includes the following contents:

[0042] S801. Assess the geological hazard risk of the study area using different high-risk assessment unit detection methods, and visualize the assessment results; further, the high-risk assessment unit detection methods include but are not limited to machine learning, information method, AHP, quotient method, and determination coefficient method;

[0043] S802. Overlay and comprehensively compare the risk analysis results of all evaluation units in the small watershed area to obtain the graded target areas for geological disaster monitoring in the area. Areas with 3 or more types of high risk are determined as the preferred monitoring target areas, areas with 1-2 types of high risk are determined as secondary monitoring target areas, and the rest are non-target areas.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] (1) The method of the present invention combines geospatial data, data mining methods, machine learning algorithms, numerical simulation methods, and InSAR technology to comprehensively evaluate potential disaster points in a region, and has the advantages of high accuracy and strong scalability. It not only solves the problems of insufficient comprehensiveness and low prediction accuracy of traditional methods that rely on single empirical judgments and historical cases, but also provides scientific and reliable technical support for the prediction and prevention of high-risk areas in the region. The quantitative evaluation method of high-risk areas in the region proposed by the present invention can provide important decision-making basis for disaster warning, engineering construction and emergency management, and minimize the social and economic losses caused by geological disasters.

[0046] (2) The method of the present invention is based on refined grid cells, combined with machine learning algorithms, numerical simulation methods, and InSAR technology for mutual verification, which significantly improves the accuracy of target area selection for surface geological disaster monitoring and realizes the quantitative assessment of high-risk areas in the surface.

[0047] (3) The method of the present invention collects historical geological disaster data and geospatial data, combines machine learning algorithms, numerical simulation methods, and InSAR technology, and uses a multi-source result fusion decision-making mechanism to construct an intelligent discrimination system for three-level target area division. When InSAR deformation monitoring, machine learning prediction, and numerical simulation all point to high risk (intersection area), it is marked as the preferred target area; 1-2 high-risk units are determined as secondary target areas, and the rest are non-point target areas. This decision-making model based on spatial overlay analysis provides an important basis for formulating scientific and reasonable geological disaster prevention and control plans, and helps optimize resource allocation and the efficient implementation of prevention and control measures. In addition, this method can also provide scientific decision-making support for urban planning, land use, infrastructure construction and other fields, and reduce the potential disaster risks caused by inappropriate construction activities in high-risk areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, a brief introduction to the drawings involved in the embodiments is now provided. It is obvious that the drawings described below are only schematic illustrations of some embodiments of the present invention. Those skilled in the art can construct other forms of drawings based on these drawings without inventive effort.

[0049] Figure 1 This is a flow chart of a multi-source data fusion determination method for small watershed geological disaster monitoring target area proposed by the present invention;

[0050] Figure 2 This is a schematic diagram of the watershed unit division proposed in Example 1 of the present invention;

[0051] Figure 3 Schematic diagram of three risk assessment results proposed in Example 1 of the present invention. DETAILED DESCRIPTION

[0052] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0053] The present invention proposes a method for determining the target area for monitoring geological disasters in small watersheds by fusing multi-source data. The method uses a machine learning algorithm to evaluate the risk of the study area; adopts a numerical simulation method to model the study area and analyze the disaster risk; uses satellites to obtain multi-phase SAR images of the study area; sets the deformation rate and selects high-risk areas for disasters. The results of the small watershed area analysis are comprehensively compared. Areas with 3 or more high-risk areas are determined as the preferred monitoring target areas, areas with 1-2 high-risk areas are determined as secondary monitoring target areas, and the rest are non-target areas. The present invention comprehensively integrates a variety of disaster point monitoring methods, greatly improving the scientificity and economy of target area selection. The multi-source data fusion determination method for monitoring target areas in small watersheds proposed by the present invention is described below in conjunction with relevant drawings and specific examples. The specific content is as follows.

[0054] Example 1:

[0055] See also Figure 1 The present invention provides a method for determining a target area for geological disaster monitoring in a small watershed based on multi-source data fusion, comprising the following steps:

[0056] Step 1: Obtain historical geological disaster data and geospatial data of the study area and establish a database; the details are as follows:

[0057] The historical geological disaster data and geospatial data of the study area include but are not limited to: disaster point coordinate data, elevation data (DEM), slope, aspect, slope shape, slope position, slope variability, aspect variability, micro-topography, profile curvature, plane curvature, land use type, slope type, terrain relief, terrain roughness, surface cutting depth, elevation variation coefficient, terrain moisture index, groundwater type, water flow dynamics index, sediment transport index, runoff modulus, distance from river, river network density, distance from fault, stratigraphic age, normalized difference vegetation index, multi-year average rainfall, POI kernel density, and distance from road.

[0058] Step 2: Construct refined evaluation units (grid units or slope units) for the small watershed study area, extract the geospatial data obtained in S1 to the corresponding grid units, and establish a dataset for quantitative evaluation of the study area; the details are as follows:

[0059] Grid cells of varying resolutions (10m, 20m, 30m, etc.) were created within the study area using ARCGIS software. Slope units were then divided using r.slopeunits. Geospatial data was extracted into evaluation units, and a machine learning database was constructed. Each grid cell contained corresponding geographic factor data to support subsequent evaluations.

[0060] Step 3: Use machine learning algorithms to conduct risk assessment on the study area and obtain the risk distribution of the study area; the details are as follows:

[0061] Use machine learning algorithms (including but not limited to tree-based models, support vector machines, artificial neural networks, and gradient boosting frameworks, specifically in the form of decision trees, random forests, support vector machines (SVMs), artificial neural networks (ANNs), XGBoost, or LightGBM) to build a hazard assessment model (using one or a combination of two or more machine learning algorithms); use disaster site data as positive samples, and randomly select an equal amount of non-disaster site data as negative samples, or through an optimized negative sample strategy, to build a training data set for the hazard assessment model for model training;

[0062] Eliminate redundant factors based on factor screening methods (including but not limited to causal analysis, correlation analysis, principal component analysis, recursive feature elimination, and geographic factor detectors, where the factor screening operation uses one or a combination of two or more of the above factor screening methods), and optimize the model's hyperparameters (including but not limited to Bayesian optimization, gradient descent algorithm, grid search, random search, and genetic algorithm, where the optimization operation uses one or a combination of two or more of the above optimization methods);

[0063] S303. Based on the risk assessment factor data set, various risk assessment factors in the data set are input into the optimized risk assessment model to obtain risk mapping.

[0064] Step 4: Obtain the geotechnical parameters of the study area; details are as follows:

[0065] Use outdoor or indoor experiments to obtain geotechnical parameters of the study area. These methods include, but are not limited to, indoor experiments such as direct shear tests, triaxial tests, and ring shear tests; and outdoor experiments such as standard penetration tests (SPTs) and cone penetration tests (CPTs).

[0066] Step 5: Use numerical simulation software to calculate the stability coefficients of all evaluation units in the study area, perform spatial analysis and visualization based on the calculation results, and obtain the high-risk distribution in the study area; the details are as follows:

[0067] The study area is modeled using numerical simulation software, and the stability coefficients of all evaluation units in the study area are calculated; further, the numerical simulation software includes but is not limited to numerical simulation using Scoops3D; the stability coefficient results of the evaluation units are spatially analyzed and visualized to obtain the high-risk distribution of the study area.

[0068] Step 6: Use SAR data platforms (including but not limited to Sentinel-1, ALOS-2 / PALSAR-2, and TerraSAR-X satellites) to obtain multi-temporal SAR images of the study area.

[0069] Step 7: Set the deformation rate threshold of the study area, and take the evaluation units with deformation greater than the threshold as high-risk areas to obtain the high-risk distribution of the study area; specifically, take the following:

[0070] Image generation was performed using image registration and interferogram generation software, PS points were extracted from StaMPS, deformation rates and time series were estimated, and a deformation rate threshold for the study area was set. Meteorological data was used to correct the atmospheric phase, and deformation maps were output and superimposed on evaluation units. Units with deformation rates exceeding the threshold were extracted to obtain the high-risk distribution of the study area. Meteorological data included EAR5.

[0071] Step 8: Use multiple disaster point risk assessment methods to overlay and analyze the risk assessment results of the small watershed area assessment unit, and identify high-risk areas as the preferred monitoring target areas; the details are as follows:

[0072] Different high-risk assessment unit detection methods are used to evaluate the geological hazard risk in the study area, and the assessment results are visualized. The high-risk assessment unit detection methods include machine learning prediction, InSAR deformation monitoring, and numerical simulation analysis. The risk analysis results of all assessment units in the small watershed area are superimposed and comprehensively compared to obtain the graded target areas for surface geological hazard monitoring. Areas with 3 or more types are high-risk areas and are determined as the preferred monitoring target areas. Areas with 1-2 types are high-risk areas and are determined as secondary monitoring target areas. The rest are non-target areas.

[0073] In summary, this embodiment uses ARCGIS software to preprocess geospatial data in the study area and establish a spatial database of geological hazards in the study area. A refined grid hazard assessment model is then constructed, revealing a mechanism for selecting targets for monitoring locations in high-risk areas. The machine learning algorithm, Random Forest, is used to assess the hazard of the study area, screening high-risk areas as targets for regional geological hazard monitoring. Scoop3D numerical simulation software is then used to calculate all possible potential hazard points within the study area, performing spatial analysis and visualization to obtain the distribution of high-risk areas in the study area, which serve as targets for regional geological hazard monitoring. Multi-temporal SAR imagery of the study area is then acquired using the Sentinel-1 satellite. A deformation rate threshold is set for the study area, and points with deformations greater than the threshold are designated as hazard-prone points. The distribution of high-risk areas in the study area is then obtained, serving as targets for regional geological hazard monitoring. Finally, the above three methods are combined to comprehensively analyze the regional geological hazard risk of small watersheds using multiple hazard point hazard assessment methods, ultimately selecting high-risk areas as the preferred targets for monitoring locations. This method provides quantitative prediction results, which provides an important basis for formulating scientific and reasonable geological disaster prevention and control plans, and helps to optimize resource allocation and effectively implement prevention and control measures.

[0074] The above description is only used to help understand the method and core essence of the present invention, but the scope of protection of the present invention is not limited thereto. For those skilled in the art, equivalent replacements or modifications based on the technical solutions and inventive concepts of the present invention within the technical scope disclosed by the present invention should be included in the scope of protection of the present invention. In summary, the contents of this specification should not be understood as limiting the present invention.

Claims

1. A method for determining target areas for geological disaster monitoring in small watersheds by fusion of multi-source data, characterized in that: The following steps are involved: S1. Obtain historical geological disaster data and geospatial data of the study area and establish a database; S2. Construct refined evaluation units for the small watershed study area. The refined evaluation units include grid units or slope units. Extract the geospatial data obtained in S1 to the corresponding grid units to establish a dataset for quantitative evaluation of the study area. S3. Use machine learning algorithms to evaluate the risk of the study area and obtain the distribution of high-risk areas in the study area; S4. Obtain geotechnical parameters of the study area; S5. Use numerical simulation software to calculate the stability coefficients of all evaluation units in the study area, perform spatial analysis and visualization based on the calculation results, and obtain the distribution of high-risk areas in the study area; S6. Use the SAR data platform to obtain multi-temporal SAR images of the study area; S7. Set a deformation rate threshold for the study area, and define the evaluation units with deformation greater than the threshold as high-risk areas to obtain the high-risk distribution of the study area; S8. Use multiple disaster point hazard assessment methods to overlay and analyze the hazard assessment results of small watershed area assessment units, and identify high-risk areas as the preferred monitoring target areas.

2. The method for determining target areas for geological disaster monitoring in small watersheds by fusion of multi-source data according to claim 1, characterized in that: The historical geological disaster data in S1 include landslide point data or surface data; the geographic spatial data include topographic factors, geological structure factors, hydrological factors, vegetation factors, meteorological factors, and human activity factors.

3. The method for determining target areas for geological disaster monitoring in small watersheds by fusion of multi-source data according to claim 2, characterized in that: The regional refined evaluation unit described in S2 was constructed using ARCGIS software.

4. The method for determining target areas for geological disaster monitoring in small watersheds by fusion of multi-source data according to claim 3, characterized in that: The S3 specifically includes the following contents: S301. Constructing a risk assessment model using a machine learning algorithm; using disaster site data as positive samples and randomly selecting an equal amount of non-disaster site data as negative samples, or by optimizing a negative sample strategy, to construct a training data set for the risk assessment model for model training; wherein the risk assessment model is constructed using one or a combination of two or more machine learning algorithms; S302. Perform factor screening based on a factor screening method to eliminate redundant factors and optimize the model's hyperparameters. S303. Based on the risk assessment factor data set, various risk assessment factors in the data set are input into the optimized risk assessment model to obtain risk mapping and obtain the distribution of high risk assessment units in the study area.

5. The method for determining target areas for geological disaster monitoring in small watersheds by fusion of multi-source data according to claim 4, characterized in that: The geotechnical parameters of the study area described in S4 include internal friction angle, cohesion, density, pore water pressure coefficient, and specific gravity. The acquisition methods include indoor experiments and outdoor experiments. Indoor tests include direct shear test, triaxial test, and ring shear test; outdoor experiments include standard penetration test (SPT) and static penetration test (CPT).

6. The method for determining target areas for geological disaster monitoring in small watersheds by fusion of multi-source data according to claim 5, characterized in that: The S5 specifically includes the following contents: S501, using numerical simulation software to model the study area and calculate the stability coefficients of all evaluation units in the study area; wherein the numerical simulation software includes Scoops3D; S502. Perform spatial analysis and visualization on the stability coefficient results of the evaluation unit to obtain the high-risk distribution in the study area.

7. A method for determining target areas for geological disaster monitoring in a small watershed by fusion of multi-source data according to any one of claim 6, characterized in that: The SAR data platforms mentioned in S6 include Sentinel-1, ALOS-2 / PALSAR-2, and TerraSAR-X satellites.

8. The method for determining target areas for geological disaster monitoring in small watersheds by fusion of multi-source data according to claim 7, characterized in that: The S7 specifically includes the following contents: S701. Use image registration and interferogram generation software to generate images, extract PS points in StaMPS, estimate deformation rate and time series, and set deformation rate thresholds for the study area; wherein the image registration and interferogram generation software includes Snap and GMTSAR; S702. Correct the atmospheric phase using meteorological data, output a deformation map and superimpose it on the evaluation unit, extract the units whose deformation rate exceeds the threshold to obtain the high-risk distribution of the study area; wherein the meteorological data includes EAR5.

9. The method for determining target areas for geological disaster monitoring in small watersheds by fusion of multi-source data according to claim 8, characterized in that: The specific S8 details are as follows: S801. Assess the geological hazard risk of the study area using different high-risk assessment unit detection methods and visualize the assessment results; wherein the high-risk assessment unit detection methods include one or more of machine learning prediction, information quantity method, hierarchical analysis method, commercial weight method, and determination coefficient method; S802. Overlay and comprehensively compare the risk analysis results of all evaluation units in the small watershed area to obtain the graded target areas for geological disaster monitoring in the area. Areas with 3 or more types are high-risk areas and are determined as the preferred monitoring target areas. Areas with 1-2 types are high-risk areas and are determined as secondary monitoring target areas. The rest are non-target areas.

Images