Comprehensive remote sensing identification method for hidden geological hazards

Abstract

The present invention relates to a method for comprehensive remote sensing identification of concealed geological disasters, specifically, InSAR+high-resolution remote sensing is used to comprehensively identify concealed geological disasters, and spatial identification and prediction of prone slopes and risky slopes are added, that is, through high-resolution remote sensing data, radar data, DEM data, stratigraphic lithology data, geological structure data, geological disaster data, landform type data, and human engineering activity data, a developed spatial identification and analysis platform is used to build a geological disaster-prone slope-risky slope spatial prediction system, so as to achieve accurate prediction of hidden geological disaster-prone sections and risky sections, predict slopes that pose greater threats to residential areas, important engineering facilities, etc. through threshold intervals set by spatial identification, comprehensively judge the identification results to delineate key control slopes, and find out hidden geological disasters and their prone slopes and risky slopes to the greatest extent, so as to reduce the loss of life and property caused by such disasters to the people.

Description

Technical Field

[0001] The invention relates to a geological disaster prediction method, in particular to a hidden geological disaster comprehensive remote sensing identification method. Background Art

[0002] The prominent characteristics of hidden geological disasters are manifested in three aspects: first, they are highly concealed. In the case of dense vegetation in mountainous areas of my country, traditional remote sensing identification methods are difficult to effectively identify signs of slope deformation and disaster signs; second, they are highly sudden. The geological background conditions in the areas where hidden geological disasters occur are usually unstable, which will cause the disaster to shift or be pulled, leading to large-scale disasters; third, they are highly destructive. The locations of hidden geological disasters are high in the mountains and steep in the slopes, which are inaccessible to human power. Large-scale geological disasters are easily transformed into mudslides, debris flows or secondary disasters after collisions and disintegrations along the way, causing the movement distance to increase to thousands of meters, greatly increasing the scope of destruction, and further amplifying the disaster by orders of magnitude, seriously threatening people’s lives and property safety.

[0003] According to the current research status at home and abroad, the traditional remote sensing technology is particularly effective in identifying explicit geological disasters, but there are still many problems to be solved in identifying hidden geological disasters. For example, the use of traditional optical remote sensing + InSAR technology to identify geological disaster bodies is not perfect, and the disaster-prone slopes and risk slopes have not been systematically integrated into the comprehensive remote sensing identification system, resulting in the identification target not being comprehensive and clear, and the weight of each indicator has not been distinguished. That is, the current research has not systematically carried out comprehensive remote sensing identification that integrates "hidden disaster bodies + prone slopes + risk slopes". Therefore, how to accurately find hidden geological disaster bodies and their prone slopes and risk slopes, establish an integrated identification system, and complete susceptibility and risk assessments are urgent problems that need to be solved, and are of great significance to the investigation and evaluation of geological disasters. Summary of the invention

[0004] The purpose of the present invention is to provide a comprehensive remote sensing identification method for concealed geological disasters to solve the problem that the existing remote sensing technology is not comprehensive and clear enough in identifying concealed geological disasters.

[0005] The object of the present invention is achieved in that:

[0006] A method for comprehensive remote sensing identification of hidden geological hazards comprises the following steps:

[0007] S1. Obtain basic geological environment data including high-resolution remote sensing data, radar data, DEM data, stratigraphic lithology data, geological structure data, geological disaster data, landform type data, and human engineering activity data in the distribution areas of hidden geological hazards.

[0008] S2. Carry out SBAS-InSAR differential interferometry measurement on each disaster slope unit in the hidden geological disaster distribution area, and perform image registration on Sentinel 01 data based on DEM data to weaken the spatial decorrelation effect between multi-scene Sentinel images, so as to form a differential interferometry phase map of the hidden geological disaster distribution area; remove the flat ground effect from the differential interferometry phase map so that it can intuitively reflect the terrain changes in the geological disaster distribution area; perform orbit refinement, coherence estimation, filtering, phase unwrapping and geocoding on the differential interferometry phase map to invert the ground target height in the disaster distribution area and remove the surface phase related to the terrain; use the filtering method to perform atmospheric delay phase correction to improve the accuracy of deformation estimation in the disaster distribution area; through the deformation extraction results of the disaster distribution area, complete the delineation of the deformation area of ​​the hidden geological disaster distribution area, and form an InSAR deformation rate map of the hidden geological disaster distribution area.

[0009] S3. Utilize high-resolution remote sensing data in basic geological environment data, combined with DEM data, to conduct three-dimensional optical remote sensing identification of hidden geological disaster distribution areas and obtain three-dimensional optical remote sensing identification data; perform data processing on the three-dimensional optical remote sensing identification data, including data correction, data preprocessing and disaster enhancement processing, to form an image base map and optical remote sensing identification map of the hidden geological disaster distribution area; analyze the internal structural characteristics of the geological disaster data based on the spectral characteristics of the hidden geological disaster distribution area and the three-dimensional spatial distribution characteristics, combined with the geological data of the geological disaster distribution area, and establish a three-dimensional identification mark for the hidden geological disaster distribution area based on DEM data and high-resolution remote sensing data.

[0010] S4. Based on the InSAR deformation rate map and three-dimensional optical remote sensing identification map of the hidden geological disaster distribution area, a sensitivity analysis of the hidden geological disaster distribution area is conducted, including weight analysis between geological disaster data and terrain slope, slope aspect, stratum lithology, and geological structure, extracting key identification factors of geological disaster susceptibility, and identifying and analyzing the vector data of key identification factors of geological disaster susceptibility, to obtain intermediate data including terrain slope, slope aspect, stratum lithology and geological structure of disaster-prone slopes in the hidden geological disaster distribution area.

[0011] S5. Based on the sensitivity analysis of disaster-prone areas in the distribution of hidden geological disasters, combined with the development characteristics and occurrence laws of hidden geological disasters, the relationship between the terrain slope, slope aspect, stratigraphic lithology and geological structure of disaster-prone slopes in the distribution of hidden geological disasters and the occurrence of geological disasters is analyzed, and the identification rules for disaster-prone slopes in the distribution of hidden geological disasters are constructed. The threshold range of key identification factors for geological disasters is determined to indirectly predict the spatial distribution areas prone to geological disasters.

[0012] S6. According to the identification rules of disaster-prone slopes in the distribution areas of hidden geological disasters, disaster-prone slopes are predicted, and risk slopes are predicted on this basis; according to the spatial distribution of hidden geological disaster-prone slopes, weights are assigned to the disaster-bearing bodies and the scope of the danger zone, and the risk slopes where hidden geological disasters are located are delineated to form a risk slope analysis diagram.

[0013] S7. Based on the InSAR deformation rate map of the hidden geological disaster distribution area, the deformation area within the hidden geological disaster distribution area is delineated to identify the hidden landslide hazards therein; optical remote sensing recognition technology is used to construct three-dimensional recognition signs of existing geological disasters and potential hidden landslide hazards on the risk slope vector data; through spatial recognition and analysis of prone slopes and risk slopes, the section of the slope where the disaster body is located is predicted to obtain comprehensive recognition data, and the field investigation area is delineated for field site investigation.

[0014] S8. Convert the comprehensive identification data into InSAR deformation rate maps, optical remote sensing identification maps, prone slope analysis maps and risk slope analysis maps. Combined with the topographic geological maps and slope analysis maps of the hidden geological disaster distribution areas, the field investigation contents ascertained during the field investigation, including the spatial distribution range, impact area, slope of specific slopes, slope direction and stratum lithology, are mapped to form spatial external survey data and database.

[0015] S9. The data on geological hazards, disaster susceptibility and risk levels obtained through field investigations will be entered into the system. After verification and correction, the slopes prone to hidden geological hazards and risky slopes will be finally identified, key prevention and control sections will be marked, and the comprehensive remote sensing identification of hidden geological hazards will be completed.

[0016] Furthermore, the key identification factors for disaster susceptibility include terrain slope, slope aspect, stratum lithology and geological structure.

[0017] Furthermore, in step S5, the threshold interval of the key identification factor of the disaster-prone slope is determined as follows:

[0018] S5-1 determines the location, scale and scope of hidden geological disasters based on high-resolution remote sensing data, radar data and DEM data of different periods in the distribution area of ​​hidden geological disasters, and then completes the human-computer interactive interpretation of the geological disasters themselves. The existing geological disasters are used to superimpose the terrain slope, slope aspect, stratigraphic lithology and geological structure of the disaster-prone slopes, so as to determine the threshold range of the disaster-prone slopes.

[0019] S5-2 Based on the existing geological disasters, past rainfall conditions and the distribution of geological disasters, the existing geological disasters and key identification factors of disaster-prone areas are input into the spatial analysis module of the ArcGIS platform for overlay analysis to obtain the slope value, slope aspect range, lithology of the collapse-prone and landslide-prone strata and disaster-controlling structures of the disaster-prone slopes.

[0020] Furthermore, the disaster-prone body includes population, buildings, transportation facilities and key projects; the scale of the disaster-prone body is determined by a neural network model, and the specific determination method is:

[0021] S6-1 obtains data samples including the number of threatened people, road damage conditions and the scale of disaster-prone bodies in the prediction area; the number of threatened people and road damage conditions in the prediction area are used as input data, and the scale of disaster-prone bodies in the prediction area is used as output data.

[0022] S6-2 normalizes the input data and divides the data samples into training set and test set in a ratio of 8:2.

[0023] S6-3 inputs the training set into the neural network of each type of disaster-prone body for training, and obtains a mathematical model for geological disaster risk assessment and prediction.

[0024] S6-4 inputs the number of threatened people and road damage in the forecast area into the forecast mathematical model to obtain the scale of the disaster-prone body in the forecast area.

[0025] The present invention utilizes InSAR+high-resolution remote sensing to comprehensively identify hidden geological disasters, and adds spatial identification and prediction of prone slopes and risky slopes, that is, through high-resolution remote sensing data, radar data, DEM data, stratigraphic lithology data, geological structure data, geological disaster data, landform type data, and human engineering activity data, through the developed spatial identification and analysis platform, a geological disaster-prone slope-risky slope spatial prediction system is constructed, thereby replacing the traditional identification method, and realizing the accurate prediction of hidden geological disaster-prone sections and risky sections, and through the threshold interval set by spatial identification, predicting the slopes that pose greater threats to residential areas, important engineering facilities, etc., and comprehensively judging the identification results to delineate key prevention and control slopes, thereby maximally finding hidden geological disasters and their prone slopes and risky slopes, so as to minimize the loss of life and property caused by such disasters to the people.

[0026] The present invention can reflect the intrinsic connection and influencing mechanism between the hidden disaster body itself and the disaster-prone slopes and risky slopes; it comprehensively covers all contents of remote sensing identification of hidden geological disasters, and the composition of the comprehensive remote sensing identification method system of hidden geological disasters is more complete, which deeply reflects the essence of identification and evaluation of hidden geological disasters, and can simultaneously compare and analyze the intrinsic connection between the identification of the disaster body itself and the identification of prone slopes and risky slopes, as well as the impact on the subsequent hazard assessment decision-making. It is scientific and reasonable, has strong applicability, high engineering application value and good effect.

[0027] This invention breaks through the traditional technical thinking of geological disaster body identification, and has completed the spatial prediction technology of "hazard body-prone slope-risk slope" based on multi-source remote sensing data. This technology uses a three-dimensional spatial superposition analysis method, and according to the spatial relationship between geological hazards and factors such as stratum lithology, geological structure, terrain slope, slope direction, and the analysis of the proneness and sensitivity characteristics and risk level, it constructs a geological disaster proneness and risk threshold prediction model, and achieves the purpose of rapid prediction of hidden geological disaster proneness and risk sections, laying the foundation for further targeted investigation and evaluation of hidden geological disasters. DETAILED DESCRIPTION

[0028] The present invention is further described below.

[0029] The comprehensive remote sensing identification method for hidden geological hazards of the present invention comprises the following steps:

[0030] S1. Obtain basic environmental data of hidden geological disaster distribution areas through data collection, platform download and other means. The basic environmental data include high-resolution remote sensing data, radar data, DEM data, stratigraphic lithology data, geological structure data, geological disaster data, landform type data and human engineering activity data.

[0031] S2. Conduct SBAS-InSAR differential interferometry on each disaster slope unit in the hidden geological disaster distribution area, and perform image registration on Sentinel 01 data based on DEM data to weaken the spatial correlation effect between multi-view Sentinel images and form a differential interferometry phase map of the geological disaster distribution area. Remove the flat ground effect of the differential interferometry phase map so that it can intuitively reflect the terrain changes in the geological disaster distribution area. Perform orbit refinement, coherence estimation, filtering, phase unwrapping and geocoding on the differential interferometry phase map to invert the ground target height in the disaster distribution area and remove the surface phase related to the terrain. Use the filtering method to perform atmospheric delay phase correction to improve the accuracy of deformation estimation in the hidden geological disaster distribution area. Through the deformation extraction results of the disaster distribution area, the deformation area of ​​the hidden geological disaster distribution area is delineated to form an InSAR deformation rate map of the hidden geological disaster distribution area.

[0032] S3. Use high-resolution remote sensing data in the basic environmental data of the hidden geological disaster distribution area, combined with DEM data, to perform three-dimensional optical remote sensing identification of the hidden geological disaster distribution area to obtain three-dimensional optical remote sensing identification data. Perform data processing on the three-dimensional optical remote sensing identification data, including data correction, data preprocessing and disaster enhancement processing, to form an image base map and optical remote sensing identification map of the hidden geological disaster distribution area. According to the spectral characteristics and three-dimensional spatial distribution characteristics of the hidden geological disaster distribution area itself, combined with the geological data of the hidden geological disaster distribution area, analyze the internal structural characteristics of the geological disaster data, and establish a three-dimensional identification mark for the hidden geological disaster distribution area based on DEM data and high-resolution remote sensing data.

[0033] S4. Comprehensively analyze the InSAR deformation rate map and three-dimensional optical remote sensing identification map of the hidden geological disaster distribution area to conduct disaster-prone sensitivity analysis on the hidden geological disaster distribution area. The specific method of disaster-prone sensitivity analysis is: through the weight analysis between geological disaster data and terrain slope, slope aspect, stratum lithology and geological structure, extract the key identification factors of disaster-proneness, and perform identification analysis on the vector data of the key identification factors of disaster-proneness, and obtain the intermediate data such as terrain slope, slope aspect, HillShade, stratum lithology and geological structure of the disaster-prone slope in the hidden geological disaster distribution area. The key identification factors of disaster-proneness include terrain slope, slope aspect, stratum lithology and geological structure.

[0034] S5. Based on the sensitivity analysis of disaster-prone areas in the distribution of hidden geological disasters, combined with the development characteristics and occurrence laws of hidden geological disasters, the relationship between the terrain slope, slope aspect, stratigraphic lithology and geological structure of disaster-prone slopes in the distribution of hidden geological disasters and the occurrence of geological disasters is analyzed, and the identification rules for disaster-prone slopes in the distribution of hidden geological disasters are constructed. The threshold range of key identification factors for disaster-proneness is determined to indirectly predict the spatial distribution areas prone to geological disasters.

[0035] The threshold intervals of the key identification factors for disaster-prone slopes are determined as follows:

[0036] S5-1 determines the location, scale and scope of hidden geological disasters based on high-resolution remote sensing data, radar data and DEM data of multiple periods of hidden geological disaster distribution areas, and then completes the human-computer interactive interpretation of the geological disasters themselves. The existing geological disasters are used to superimpose the terrain slope, slope aspect, stratigraphic lithology and geological structure of disaster-prone slopes, thereby determining the threshold range of disaster-prone slopes.

[0037] S5-2 Based on the existing geological disasters, past rainfall conditions and the distribution of geological disasters, the existing geological disasters and key identification factors of disaster-prone areas are input into the spatial analysis module of the ArcGIS platform for overlay analysis to obtain the slope value, slope aspect range, lithology of the collapse-prone and landslide-prone strata and disaster-controlling structures of the disaster-prone slopes.

[0038] S6. Based on the identification rules of disaster-prone slopes in the distribution areas of hidden geological disasters, the disaster-prone slopes are predicted, and on this basis, the risk slopes are predicted. According to the spatial distribution of disaster-prone slopes, the population, buildings, transportation facilities, key projects and other disaster-prone objects and the scope of the danger zone obtained by land survey data are weighted, and the risk slopes where hidden geological disasters are located are delineated to form a risk slope analysis map for field investigation and comprehensive evaluation.

[0039] The scale of the disaster-prone body can be determined by the neural network model. The specific method is:

[0040] S6-1 obtains data samples of the prediction area, the data samples include the number of people threatened, road damage conditions and the scale of disaster-prone objects in the prediction area; and uses the number of people threatened and road damage conditions in the prediction area as input data, and uses the scale of disaster-prone objects in the prediction area as output data.

[0041] S6-2 normalizes the input data and divides the data samples into training set and test set in a ratio of 8:2.

[0042] S6-3 inputs the training set into the neural network of each type of disaster-prone body for training, and obtains a mathematical model for geological disaster risk assessment and prediction.

[0043] S6-4 inputs the number of threatened people and road damage in the forecast area into the forecast mathematical model to obtain the scale of the disaster-prone body in the forecast area.

[0044] S7. Based on the InSAR deformation rate map of the hidden geological disaster distribution area, the deformation area within the hidden geological disaster distribution area is delineated to identify the hidden landslide hazards therein. Using optical remote sensing recognition technology, a three-dimensional identification mark of the geological disasters that have occurred and the potential hidden landslide hazards is constructed on the risk slope vector data. Through the spatial recognition and analysis of the prone slope and risk slope, the section of the slope where the disaster body is located is predicted, and the comprehensive recognition data is obtained. The field investigation area is delineated and the field site investigation is carried out.

[0045] S8. Convert the obtained comprehensive identification data into InSAR deformation rate maps, optical remote sensing identification maps, prone slope analysis maps and risk slope analysis maps. Combined with the topographic geological maps and slope analysis maps of hidden geological disaster distribution areas collected by the natural resources department, the spatial distribution range, impact area, slope, slope direction and stratigraphic lithology of specific slopes identified in the field investigation will be mapped to form spatial external survey data and database.

[0046] The method of converting the comprehensive identification data into maps is: based on the ArcGIS map output platform, superimpose the InSAR deformation rate, optical remote sensing identification results, prone slope analysis data, risk slope analysis, etc., and decorate the data into the result map view mode, and output them as InSAR deformation rate maps, optical remote sensing identification maps, prone slope analysis maps and risk slope analysis maps.

[0047] S9. The data on geological hazards, disaster susceptibility and risk levels obtained through field investigations are entered into the system and reviewed and corrected. Finally, the slopes prone to hidden geological hazards and risky slopes are identified, and key prevention and control sections are marked, thus completing the comprehensive remote sensing identification of hidden geological hazards.

[0048] In the comprehensive remote sensing identification of hidden geological hazards in a county in Yunnan Province, after obtaining the basic environmental data of the area, the ArcGIS spatial identification and analysis platform and GAMMA software analysis function were used to select multiple periods of Sentinel 01 data, GF-2 data, 30m DEM data, supplemented by terrain data, 1:50000 regional geological background data, and detailed geological disaster survey data to build a comprehensive remote sensing identification and analysis platform GAMMA for the area; ArcGIS10.6 and GAMMA platforms were used to conduct time-series InSAR differential interferometry in the area, obtain the InSAR deformation rate of the ascending and descending orbits in the area, and delineate the larger deformation area. Combined with the meter-level GF-2 data covering the county and the DEM elevation model, the morphological characteristics were analyzed through the ArcGlobe three-dimensional visualization function, the boundaries of hidden landslide hazards were delineated, and vectorized data were formed, which were then superimposed on the ArcMap analysis platform to form the InSAR deformation rate map and landslide hazard optical remote sensing identification map of the area. The ArcGIS spatial identification and analysis platform was used to determine the slope, slope direction, stratum lithology and geological structure of the disaster-prone slopes. Through sensitivity analysis, the critical slope of the disaster-prone slopes in the county was set at 25°~55°, the critical slope direction interval was set at SE, S and SW, and the lithology was set at loose accumulation layer, mudstone and shale. Through the above steps, a total of 142 hidden landslide hazards and 236 disaster-prone slopes were identified, covering an area of ​​406.9km 2 Based on the comprehensive remote sensing identification results, a key survey area of ​​582.9 km 2, accounting for 32.3% of the total area of ​​the county. After field investigation and mapping the field investigation content, spatial external survey data and database are formed. The data obtained through the field investigation is entered into the system, reviewed and corrected, and finally 127 landslide hazards located in the key investigation area are determined, accounting for 72.6%, and 194 risk slopes located in the key investigation area are determined, accounting for 82.3%. The comprehensive remote sensing identification of hidden geological disasters in the county has been completed. Through ROC verification, the accuracy rate of identifying landslides using the remote sensing identification method of the present invention reaches 80.6%.

Claims

1. A method for comprehensive remote sensing identification of hidden geological hazards, characterized in that: The following steps are involved: S1. Obtain basic geological environment data including high-resolution remote sensing data, radar data, DEM data, stratigraphic lithology data, geological structure data, geological disaster data, landform type data, and human engineering activity data in the distribution area of ​​hidden geological hazards; S2. Conduct SBAS-InSAR differential interferometry measurement on each disaster slope unit in the hidden geological disaster distribution area, perform image registration on Sentinel 01 data based on DEM data, and form a differential interferometry phase map of the hidden geological disaster distribution area; perform orbit refinement, coherence estimation, filtering, phase unwrapping and geocoding processing on the differential interferometry phase map, invert the ground target height in the disaster distribution area, and remove the surface phase related to the terrain; The atmospheric delay phase correction is performed using the filtering method; the deformation area of ​​the hidden geological disaster distribution area is delineated through the deformation extraction results of the disaster distribution area, and the InSAR deformation rate map of the hidden geological disaster distribution area is formed; S3. Use high-resolution remote sensing data in basic geological environment data, combined with DEM data, to perform three-dimensional optical remote sensing identification of hidden geological disaster distribution areas to obtain three-dimensional optical remote sensing identification data; perform data processing on the three-dimensional optical remote sensing identification data, including data correction, data preprocessing and disaster enhancement processing, to form an image base map and optical remote sensing identification map of the hidden geological disaster distribution area; analyze the internal structure characteristics of the geological disaster data based on the spectral characteristics and three-dimensional spatial distribution characteristics of the hidden geological disaster distribution area itself, combined with the geological data of the geological disaster distribution area, and establish a three-dimensional identification mark for the hidden geological disaster distribution area based on DEM data and high-resolution remote sensing data; S4. Comprehensively analyze the InSAR deformation rate map and three-dimensional optical remote sensing identification map of the hidden geological disaster distribution area, and conduct a sensitivity analysis of the hidden geological disaster distribution area, including weight analysis between geological disaster data and terrain slope, slope aspect, stratum lithology, and geological structure, extract key identification factors of geological disasters, and conduct identification analysis on the vector data of key identification factors of geological disasters, and obtain intermediate data including terrain slope, slope aspect, stratum lithology and geological structure of disaster-prone slopes in the hidden geological disaster distribution area; S5. Based on the sensitivity analysis of hidden geological disaster distribution areas, combined with the development characteristics and occurrence laws of hidden geological disasters, the relationship between the terrain slope, slope aspect, stratum lithology and geological structure of the disaster-prone slopes in the hidden geological disaster distribution areas and the occurrence of geological disasters is analyzed, and the identification rules of the disaster-prone slopes in the hidden geological disaster distribution areas are constructed. The threshold range of the key identification factors of geological disasters is determined to indirectly predict the spatial distribution areas prone to geological disasters; S6. According to the identification rules of the disaster-prone slopes in the hidden geological disaster distribution area, the disaster-prone slopes are predicted, and on this basis, the risk slopes are predicted; according to the spatial distribution of the hidden geological disaster-prone slopes, the weights are assigned to the disaster-bearing bodies and the range of the danger zone, and the risk slopes where the hidden geological disasters are located are delineated to form a risk slope analysis diagram; S7. Based on the InSAR deformation rate map of the hidden geological disaster distribution area, the deformation area within the hidden geological disaster distribution area is delineated to identify the hidden landslide hazards therein; the three-dimensional identification marks of the geological disasters that have occurred and the potential hidden landslide hazards are constructed on the risk slope vector data using optical remote sensing recognition technology; through the spatial recognition analysis of prone slopes and risk slopes, the section of the slope where the disaster body is located is predicted to obtain comprehensive recognition data, and the field investigation area is delineated for field site investigation; S8. Convert the comprehensive identification data into InSAR deformation rate maps, optical remote sensing identification maps, prone slope analysis maps and risk slope analysis maps, combine them with the topographic geological map and slope analysis map of the hidden geological hazard distribution area, and draw the field survey contents including the spatial distribution range, impact area, slope, slope direction and stratum lithology of the specific slope found out in the field survey to form spatial external survey data and database; S9. The data on geological hazards, disaster susceptibility and risk levels obtained through field investigations will be entered into the system. After verification and correction, the slopes prone to hidden geological hazards and risky slopes will be finally identified, key prevention and control sections will be marked, and the comprehensive remote sensing identification of hidden geological hazards will be completed.

2. The method for comprehensive remote sensing identification of hidden geological hazards according to claim 1 is characterized in that: The key identification factors for disaster susceptibility include terrain slope, slope aspect, stratum lithology and geological structure.

3. The method for comprehensive remote sensing identification of hidden geological hazards according to claim 2 is characterized in that: In step S5, the threshold interval of the key identification factor of the disaster-prone slope is determined as follows: S5-1 Determine the location, scale and scope of hidden geological disasters based on high-resolution remote sensing data, radar data and DEM data of different periods in the distribution area of ​​hidden geological disasters, and then complete the human-computer interactive interpretation of the geological disasters themselves. The geological disasters that have been formed are used to superimpose the terrain slope, slope direction, stratum lithology and geological structure of the slopes prone to disasters, so as to determine the threshold range of the slopes prone to disasters; S5-2 Based on the existing geological disasters, past rainfall conditions and the distribution of geological disasters, the existing geological disasters and key identification factors of disaster-prone areas are input into the spatial analysis module of the ArcGIS platform for overlay analysis to obtain the slope value, slope aspect range, lithology of the collapse-prone and landslide-prone strata and disaster-controlling structures of the disaster-prone slopes.

4. The method for comprehensive remote sensing identification of hidden geological hazards according to claim 1 is characterized in that: The disaster-prone bodies include population, buildings, transportation facilities and key projects; the scale of the disaster-prone bodies is determined by a neural network model, and the specific determination method is: S6-1 obtains data samples including the number of people threatened, road damage and the scale of disaster-prone bodies in the forecast area; the number of people threatened and road damage in the forecast area are used as input data, and the scale of disaster-prone bodies in the forecast area is used as output data; S6-2 normalizes the input data and divides the data samples into training set and test set in a ratio of 8:2; S6-3 inputs the training set into the neural network of each type of disaster-prone body for training, and obtains a mathematical model for geological disaster risk assessment and prediction; S6-4 inputs the number of threatened people and road damage in the forecast area into the forecast mathematical model to obtain the scale of the disaster-prone body in the forecast area.