承灾体信息提取及空间分布研究的不足已成为制约自然灾害风险评估发展的瓶颈之一。建立快速准确获取承灾体信息及其空间分布的方法与模型，是使风险评估方法更加完善，评估结果更加精细可靠从而更加具有应用价值的关键，因此，承灾体信息提取及空间分布模拟研究对于提高区域自然灾害风险防范与管理水平有着重要意义。项目拟基于多源遥感数据，运用人工智能、盲信号分离、面向对象分析等方法，对人口和建筑这两类重要承灾体进行信息提取与分布模拟研究。对于建筑物承灾体，拟结合高分辨率遥感影像与立体像对数据，运用面向对象方法和立体测量原理等进行三维重建，并结合其它辅助信息确定用途、结构类型等关键属性；对于人口承灾体，将基于夜间灯光强度、不透水面丰度、建筑信息等估算因子，结合统计数据，采用粒子群优化方法、多智能体模型等建立空间分布模拟模型。研究成果将有助于提高承灾体暴露性评价的精度，为实现高精度灾害风险评估提供方法和数据支持。

Spatial distribution of exposed elements is as important as vulnerability, and should be taken into consideration when risk assessment is performed. Detailed and accurate mapping of exposed elements is the key of producing fine and realistic risk assessment results, which, however, is seldom performed in existing risk assessment activities. In this project, we will use intelligent optimization methods to simulate spatial distribution of population and buildings, based on spatial information extracted from multi-source Remote Sensing data. High resolution images and stereo image will be used to extract three-dimensional information of buildings. As for population, we will use intelligent optimization method and multi-agent model to simulate its spatial distribution based on nighttime light intensity, impervious coverage information and three-dimensional information of buildings respectively. The results can be used to improve exposure assessment and furthermore realize high resolution risk assessment of natural hazards.