Research Of Remote Sensing Application Technology Based On Earthquake Disaster Assessment

There are multiple methods to obtain the region and loss after seismic disaster using Remote Sensing. The most and important methods are to extract seismic disaster by detecting change using pre-image and post-image or classifying post-image. The precision of change detection is restricted by geometric correction and registration method. As the development of Remote Sensing, change detection and classification technique are supposed to be developed in order to extract the information of seismic disaster faster and more precise.In this paper, SPOT images of Jiashi earthquake, which took place in 2003, were warped using the way of non-ground control points, and got better precision. IRS images of Jiji earthquake, which took place in Taiwan in 1999, were registered by employing Automatic registration of ENVI 4.2.Besides, the registration precision achieved 0.5 pixel with little manual-operation which accorded with change detection. The loss of disaster area in Jiji was acquired fast by the analysis of sub-structure change and the result of analysis accorded with investigation. The local damaged buildings were extracted exactly by processing the post-earthquake aerial photographs of Tangshan and Ikonos images of Puji in India , based on the features of gray, texture, scale and shape and using the methods of multi-scale segment and object-based image analysis. The compared result proved that the methods of multi-scale segment and object-based image analysis were better than conventional classification. Base on ERDAS, several models were built through processing and analyzing seismic images using EML, SML and C Tookit, such as the building shadow extracting model, building extracting model, pseudo color synthetic model, PCA model based on texture feature. Images of typical seismic disaster area were also processed using above models and better production were gotten.On the first hand, the emphasis in this paper is that using non-ground control points and object-base image analysis evaluated the seismic damage, which provided preference to other researchers using RS to evaluate seismic damage. On the other hand, the models based on popular RS software provided basis for further research on seismic image processing.