基于同异源高分辨率遥感立体影像的震害房屋倒塌三维精细化评估

基于高分辨率遥感影像的震害房屋倒塌快速评估是减灾救援的重大需求，是震害损失实物量评估的重要组成部分。为了提高房屋倒塌评估的精度，同时针对地震发生期间不容易获得同源立体影像的问题，提出利用同异源高分辨率遥感立体影像（本项目的同异源遥感立体影像由侧视合成孔径雷达影像，光学影像，无人机影像组成的立体影像）进行震害房屋倒塌三维精细化评估方法。该方法首先构建同异源遥感立体定位框架，通过揭示同异源定位误差的规律、提出震区同异源遥感定位精度提高模型，通过构建震前、震后精细数字地表模型、提取高精度的震害房屋倒塌三维信息，为快速实物量精细化评估提供理论基础和技术指导。

Rapid assessment of earthquake collapsed buildings based on high resolution sensor imagery (HRSI), as a great requirement of earthquake relief and rescue, is an important part of seismic damage loss assessment. In order to improve the accuracy of assessment of collapsed buildings, as well as to solve the problem of not easily receiving homologous stereo images, this paper puts forward a three-dimensional accuracy assessment method of collapsed buildings using HRSI (the different sources stereo imagery in this project is formed by side-looking SAR image, optical image, and unmanned aerial vehicle image of stereo images). This approach first builds a stereo geo-positioning framework of homologous sensors and heterologous sensors; by revealing the laws of geo-positioning errors between homologous and heterologous stereo images, it puts forward the bias-compensation of high accuracy of HRSI in the earthquake zone, and extracts three-dimensional information of collapsed buildings by generating an high precise digital surface model before and after the earthquake, thus providing theoretical basis and technical guidance for fast converting intensification assessment.