Remote Sensing Monitoring Of Disaster-pregnant Environment Of Landslide And Debris Flow In Shigatse Area

Disaster-pregnant environment is the environmental background of disaster occurrence,which is formed by the interaction of many factors from nature and society.The improvement of disaster-pregnant environment can effectively reduce the occurrence and harm of disasters.Therefore,effective monitoring of disaster-pregnant environment has become an important subject of disaster early warning and protection.The machine learning and remote sensing technology are combined to improve the scope and accuracy of environmental monitoring.In this paper,the main study area is Shigatse area.The disaster-pregnant environmental sensitivity analysis model of landslide and debris flow is established based on historical disaster data and satellite remote sensing data and ground meteorological data.The monitoring research of disaster-pregnant environment of landslide and debris flow is completed.Through the study on the influencing factors of disaster data,the spatial and temporal changes of land cover in the area are analyzed,which provides support for remote sensing monitoring of disaster-pregnant environment of landslide and debris flow and scientific basis for disaster prevention and mitigation in the area.Firstly,the cleaning of disaster data is completed,and the spatial distribution of landslide and debris flow disasters is obtained.Based on Xgboost,the feature importance ranking is obtained.Eight most important disaster-causing factors are extracted,and the evaluation model of disaster-causing factors and disaster-pregnant environment is established.The correlation between the spatial-temporal distribution characteristics of disaster-causing factors and disaster data is quantitatively analyzed.Then,the image preprocessing of MOD09A1 is completed,and the multi-temporal remote sensing data from 2001 to 2017 are obtained.Land cover classification of remote sensing images are obtained by random forest,and the accuracy is verified by high-resolution data and MCD12Q1 products.The results show that the classification accuracy and Kappa coefficient can meet the requirements of land cover change analysis.The results can be used as reliable data sources for environmental monitoring.Finally,using remote sensing software such as ENVI and ArcGIS,the changes of various environmental factors in the study area are analyzed.Including the long-term variation of surface meteorological data,the distribution relationship between topographic factors and disaster data,the spatial and temporal changes of land cover classification data,the dangerous degree of landslide and debris flow disaster data and the environmental sensitivity of disaster-pregnant environment of landslide and debris flow.Combined with the change analysis of these environmental factors,the monitoring of disaster-pregnant environment of landslide and debris flow is completed.The results show that the disaster-pregnant environment in Shigatse needs to be given great attention.The local government should formulate reasonable land use planning,disaster prevention and mitigation policies,and take effective measures to improve the disaster-pregnant environment in the study area.