Establish high-dimensional heterogeneous spatio-temporal disaster monitoring data fusion, cleaning and consistent expression, build ultra-fine resolution spatio-temporal disaster data processing methods in affected areas, develop multi-scale heterogeneous positioning, energy based positioning, wave based positioning and other traceability positioning technologies, and realize accurate and reliable positioning of disaster sources or disaster damage with the help of big data and cloud platforms. Based on the "disaster chain" mechanism and the disaster characteristics of urban infrastructure, an accident evolution model driven by different disaster scene data and domain knowledge is constructed. Based on field simulation, GIS and XR technology, a disaster trend deduction method based on deep learning network and a situation simulation deduction system based on virtual reality integration are developed.