城市房屋震害的区域分析法--与点特征判据算法的对比研究

根据高分辨率遥感影像震害房屋的基本特征,提出在该影像条件下,震害房屋信息的自动获取技术应发展以区域分析为基础的识别算法,并给出了适于震害房屋信息获取的区域分析算法的基本概念、框架及实现过程.在此基础上,介绍了一种典型的区域分析框架下的震害房屋自动识别算法,并与点特征判据算法进行了对比.结果表明,基于图像区域分析算法的识别具有很高的精度和易于实现的客观条件.最后指出,在高分辨率卫星遥感影像将成为今后震害信息获取主要数据源的情况下,开发区域分析框架下的震害房屋信息识别算法是未来的发展方向.

Imagery region analyzing method of urban building's earthquake damage:a comparative study with point characteristics-based criterion approaches

This paper first analyzes the basic characteristics of seismic building damage in high resolution image,then points out that the seismic building damage information-automatically acquiring approaches with high resolution image should develop the identificalion algorithm based on region analysis,and presents the basic concepts,mode and primary processing of the algorithm. Besides these,a typical approach based on region analysis is presented,and is compared with the point characteristics based criterion appoaches. The results show that the approaches based on region analysis possess better precision and implementing conditions. Then a conclusion can be drawn that as high resolution image has become the main data source,the automatic identification approaches based on region analysis will be developed to adapt the requirement of change in data source.