Search

CN-121982830-A - Geological disaster early warning method based on remote sensing monitoring

CN121982830ACN 121982830 ACN121982830 ACN 121982830ACN-121982830-A

Abstract

The invention relates to the technical field of geological disaster early warning, in particular to a geological disaster early warning method based on remote sensing monitoring. The geological disaster risk dataset of the target area is constructed by collecting geological remote sensing image data, topography and topography data, hydrometeorologic data, seismic data, groundwater level data and human activity data in the target area. Secondly, according to the geological disaster risk data set of the target area, the earthquake risk score, landslide and debris flow risk score, flood risk score and ground subsidence risk score are obtained. Then, according to the earthquake risk score, the landslide and debris flow risk score, the flood risk score and the ground subsidence risk score, a geological disaster comprehensive risk score is generated through multi-parameter fusion, and a geological disaster early warning grade is obtained.

Inventors

  • WEI YIMING
  • YANG YE
  • ZHOU HAO

Assignees

  • 魏一鸣

Dates

Publication Date
20260505
Application Date
20251225

Claims (10)

  1. 1. The geological disaster early warning method based on remote sensing monitoring is characterized by comprising the following steps of: collecting geological remote sensing image data, topography and topography data, hydrological data, seismic data, groundwater level data and human activity data in a target area, and constructing a geological disaster risk data set of the target area; Obtaining a seismic risk score, a landslide and debris flow risk score, a flood risk score and a ground subsidence risk score according to the geological disaster risk dataset of the target area; inputting the geological remote sensing image data, the topography and geomorphic data and the seismic data into a seismic risk monitoring model to obtain a seismic feature vector, and generating a seismic risk score according to the seismic feature vector; generating a geological disaster comprehensive risk score through multi-parameter fusion according to the earthquake risk score, the landslide and debris flow risk score, the flood risk score and the ground subsidence risk score; And obtaining the geological disaster early warning grade according to the geological disaster comprehensive risk score.
  2. 2. The geological disaster early warning method based on remote sensing monitoring according to claim 1, wherein the geological remote sensing image data comprises geological optical remote sensing images, geological radar remote sensing images, geological multispectral data and geological hyperspectral data, the terrain and topography data comprises digital elevation models, gradient data, slope data, soil layer thickness data, surface roughness data and terrain fault data, the hydrometeorologic data comprises rainfall, river water level, evaporation capacity, soil humidity, temperature, wind speed and wind direction, the seismic data comprises time, earthquake magnitude, earthquake center position, earthquake source depth data and fault density data of historical earthquake events, the underground water level data comprises underground water level monitoring data, underground water exploitation amount data and underground water level descending rate data, and the human activity data comprises mining activity data, urban expansion data, underground engineering data, water resource utilization data, road construction data and vegetation destruction data.
  3. 3. The geological disaster early warning method based on remote sensing monitoring according to claim 1, wherein the earthquake risk monitoring model comprises an input layer, an earthquake risk feature extraction layer, an earthquake risk important feature fusion layer and an earthquake risk scoring data output layer; The input layer is used for inputting the geological optical remote sensing image, the geological radar remote sensing image, the topography and geomorphology data and the seismic data in the geological remote sensing image data into the seismic risk monitoring model, and the seismic risk feature extraction layer is used for extracting geological feature images in the geological optical remote sensing image and the geological radar remote sensing image, and the specific steps comprise: Inputting the geological optical remote sensing image and the geological radar remote sensing image into a convolution layer with the size of 16 of a convolution kernel of 3 multiplied by 3 to obtain a first geological feature image, inputting the first geological feature image into a convolution layer with the size of 32 of a convolution kernel of 3 multiplied by 3 to obtain a second geological feature image, carrying out batch normalization on the second geological feature image, transmitting the feature images after batch normalization to a Sigmoid activation function to obtain a third geological feature image, splicing the second geological feature image and the third geological feature image according to the channel dimension, inputting the second geological feature image and the third geological feature image into a convolution layer with the size of 16 of the convolution kernel of 1 multiplied by 1, and adding a channel attention module after the convolution layer to obtain a fourth geological feature image; Inputting the fourth geological feature map, the topography and relief data and the seismic data into the seismic risk important feature fusion layer to generate a seismic risk important feature vector, wherein the seismic risk important feature vector comprises surface deformation features, fault activity conditions, topography and relief features and historical seismic event features; the seismic risk score data output layer is used for outputting the seismic risk score.
  4. 4. The geological disaster warning method based on remote sensing monitoring as set forth in claim 3, wherein the calculation formula of the earthquake risk score is: Wherein R eaq is a seismic risk score, σ is a Sigmoid function, F defor is a surface deformation feature, F fa is a fault activity condition feature, F ter is a topography feature, F his is a historical seismic event feature, ω 1 is a surface deformation feature weight, ω 2 is a fault activity condition feature weight, ω 3 is a topography and historical seismic event interaction feature weight, ω 4 is a surface deformation and fault activity interaction feature weight, and b is a bias term.
  5. 5. The geological disaster early warning method based on remote sensing monitoring as set forth in claim 1, wherein the calculation process of the landslide and debris flow risk scores comprises the following steps: According to the landform and landform data and the human activity data, obtaining a feature vector influencing landslide and debris flow, and according to the feature vector, obtaining a first landslide and debris flow risk speed, wherein a calculation formula is as follows: Wherein V is the speed of landslide and debris flow, G is gravity acceleration, H is the gravity fall height of the landslide body, θ is the gradient angle, L is the gradient length, ra is the rainfall, H is the human activity characteristic, M is the soil humidity, and G c is the vegetation coverage; Generating a first landslide and debris flow risk score according to the first landslide and debris flow risk speed, wherein a calculation formula is as follows: Wherein, R init is the first landslide and debris flow risk score, and alpha 1 is the adjustment coefficient.
  6. 6. The geological disaster early warning method based on remote sensing monitoring of claim 5, wherein the second landslide and debris flow risk speeds are generated according to the seismic data and the first landslide and debris flow risk speeds in the target area, and the calculation formula is as follows: Wherein V sei is the speed of landslide and debris flow after the earthquake is influenced, V is the speed of landslide and debris flow, delta 1 is an influence coefficient, E is the magnitude, D is the depth of the earthquake focus, and F is the fault activity frequency; generating a second landslide and debris flow risk score according to the second landslide and debris flow risk speed, wherein a calculation formula is as follows: Wherein R sei is a second landslide and debris flow risk score, and alpha 2 is an adjustment coefficient; According to the first landslide and mud-rock flow risk score and the second landslide and mud-rock flow risk score, landslide and mud-rock flow risk scores are obtained, and a calculation formula is as follows: R land_mud ＝γ 1 R init +γ 2 R sei ; Wherein, R land_mud is a landslide and debris flow risk score, R init is a first landslide and debris flow risk score, γ 1 is a first landslide and debris flow risk score weight, and γ 2 is a second landslide and debris flow risk score weight.
  7. 7. The geological disaster early warning method based on remote sensing monitoring as set forth in claim 1, wherein the calculating process of the flood risk score comprises the following steps: collecting the topography and relief data, the hydrometeorologic data and the human activity data of a flood basin of a target area to construct a flood risk data set, calculating the water flow speed and the river flow of the flood basin, generating the flood risk score based on the water flow speed and the river flow, and the water flow speed calculation formula is as follows: Wherein V w is the water flow speed, n is the Manning roughness coefficient, r h is the hydraulic radius, h w is the height of the gravity center drop of the terrain gradient, and L w is the horizontal distance of the terrain gradient; the river flow calculation formula is as follows: W=c w ·r w ·a w ; Wherein W is river flow, c w is runoff coefficient, r w is rainfall intensity, and a w is confluence area; The flood risk score calculation formula is as follows: R w ＝β v V w +β w W; Wherein R w is a flood risk score, β v is a current velocity weight, V w is a current velocity, β w is a river flow weight, and W is a river flow.
  8. 8. The geological disaster early warning method based on remote sensing monitoring as set forth in claim 1, wherein the calculation process of the ground collapse risk score comprises the following steps: collecting ground collapse risk data when the ground is not collapsed, and constructing a first ground collapse risk data set, wherein the first ground collapse risk data comprises the hydrological data, the groundwater level data and the human activity data; Collecting ground collapse risk data when the ground collapses, and constructing a second ground collapse risk data set, wherein the second ground collapse risk data comprises the topography and geomorphology data, the hydrological meteorological data, the groundwater level data and the human activity data; Meanwhile, ground collapse risk data influenced by earthquakes, landslides and debris flows are collected, and a third ground collapse risk data set is constructed, wherein the third ground collapse risk data comprises the topography and topography data, the hydrological meteorological data, the earthquake data, the groundwater level data and the human activity data; according to the first ground collapse risk data set, the second ground collapse risk data set and the third ground collapse risk data set, calculating a ground collapse risk score, wherein a calculation formula is as follows: R land_ssd ＝κ 1 R 1 +κ 2 R 2 +κ 3 R 3 ; Wherein R land_ssd is a ground collapse risk score, κ 1 is a first ground collapse risk score weight, R 1 is a first ground collapse risk score data score, κ 2 is a second ground collapse risk score weight, R 2 is a second ground collapse risk score data score, κ 3 is a third ground collapse risk score weight, and R 3 is a third ground collapse risk score data score.
  9. 9. The geological disaster early warning method based on remote sensing monitoring as set forth in claim 1, wherein the calculation formula of the geological disaster comprehensive risk score comprises: R=α 1 R eaq +α 2 R sei +α 3 R w +α 4 R land_ssd ; Wherein R is a geological disaster comprehensive risk score, α 1 is a seismic risk score weight, R eaq is a seismic risk score, α 2 is a landslide and debris flow risk score weight, R sei is a landslide and debris flow risk score, α 3 is a flood risk score weight, R w is a flood risk score, α 4 is a ground collapse risk score weight, R land_ssd is a ground collapse risk score, and α 1 +α 2 +α 3 +α 4 =1.
  10. 10. The geological disaster early warning method based on remote sensing monitoring according to claim 1, wherein the geological disaster early warning comprises the following steps: Setting a geological disaster early warning threshold set TH= { p 1 ,p 2 }, wherein when the comprehensive risk score of the geological disaster is smaller than p 1 , the corresponding geological disaster early warning is low risk early warning, when the comprehensive risk score of the geological disaster is in a section [ p 1 ,p 2 ], the corresponding geological disaster early warning is medium risk early warning, and when the comprehensive risk score of the geological disaster is not smaller than p 2 , the corresponding geological disaster early warning is high risk early warning, wherein the range of the comprehensive risk score of the geological disaster is [0,1], and the values of p 1 and p 2 are 0.3 and 0.7 respectively.

Description

Geological disaster early warning method based on remote sensing monitoring Technical Field The invention relates to the technical field of geological disaster early warning, in particular to a geological disaster early warning method based on remote sensing monitoring. Background Geological disasters are natural phenomena caused by changes in the internal dynamics of the earth or the external environment, and common types include earthquakes, landslides and debris flows, floods, ground subsidence, and karst subsidence. These disasters often occur in areas where the geological structure is complex, the climate is variable, or the effects of human activity are significant. Geological disasters not only form a serious threat to the safety and economic development of human society, but also have long-term influence on the environment. Therefore, an effective monitoring and early warning system is critical to mitigate damage from these hazards. Geological disaster research and early warning are always important issues in global scientific research and engineering practice, especially for areas with frequent natural disasters. Most of traditional geological disaster early warning systems focus on monitoring of single disaster types, although research and monitoring methods of single disasters achieve certain effects in respective fields, the methods often cannot comprehensively evaluate and early warn comprehensive risks among various disaster interactions in face of complex and changeable natural environments and geological conditions. Therefore, a geological disaster early warning method based on remote sensing monitoring is provided. Disclosure of Invention The invention aims to provide a geological disaster early warning method based on remote sensing monitoring, which constructs a geological disaster risk data set of a target area by collecting geological remote sensing image data, topography and geomorphic data, hydrological meteorological data, seismic data, groundwater level data and human activity data in the target area. Secondly, according to the geological disaster risk data set of the target area, the earthquake risk score, landslide and debris flow risk score, flood risk score and ground subsidence risk score are obtained. Then, according to the earthquake risk score, the landslide and debris flow risk score, the flood risk score and the ground subsidence risk score, a geological disaster comprehensive risk score is generated through multi-parameter fusion, and a geological disaster early warning grade is obtained. In order to achieve the above purpose, the present invention provides the following technical solutions: collecting geological remote sensing image data, topography and topography data, hydrological data, seismic data, groundwater level data and human activity data in a target area, and constructing a geological disaster risk data set of the target area; Obtaining a seismic risk score, a landslide and debris flow risk score, a flood risk score and a ground subsidence risk score according to the geological disaster risk dataset of the target area; inputting the geological remote sensing image data, the topography and geomorphic data and the seismic data into a seismic risk monitoring model to obtain a seismic feature vector, and generating a seismic risk score according to the seismic feature vector; generating a geological disaster comprehensive risk score through multi-parameter fusion according to the earthquake risk score, the landslide and debris flow risk score, the flood risk score and the ground subsidence risk score; And obtaining the geological disaster early warning grade according to the geological disaster comprehensive risk score. Preferably, the geological remote sensing image data comprises a geological optical remote sensing image, a geological radar remote sensing image, geological multispectral data and geological hyperspectral data, the topography and topography data comprises a digital elevation model, gradient data, slope data, soil layer thickness data, surface roughness data and topography fault data, the hydro-meteorological data comprises rainfall, river water level, evaporation capacity, soil humidity, temperature, wind speed and wind direction, the seismic data comprises time, earthquake magnitude, earthquake center position, earthquake source depth data and fault density data of historical earthquake events, the groundwater level data comprises groundwater level monitoring data, groundwater exploitation amount data and groundwater level descending rate data, and the human activity data comprises mining activity data, city expansion data, underground engineering data, water resource utilization data, road construction data and vegetation destruction data. Preferably, the earthquake risk monitoring model comprises an input layer, an earthquake risk feature extraction layer, an earthquake risk important feature fusion layer and an earthquake risk scoring data out