Abstract: After a large scale earthquake, emergency transportation for rescue cries for road damage information in disaster area. However the precision of road damage extraction from remote sensing is general low because the method of which is based on pixels. This paper presents an object-oriented method, which uses many image features synthetically and combines with pre-earthquake’s GIS vector road to extract road information, then identifies road damage information by integrity degree of road. High resolution remote sensing of Wenchuan disaster area is implemented as examples. The result indicates that this method can improve quickness and efficiency of damage detection by comparing with visual interpretation.

Key words: high resolution remote sensing, object-oriented, road, damage detection

摘要： 大地震之后，紧急救援物资运输迫切需要了解灾区道路的震害信息，然而当前对遥感影像中道路的震害信息提取大多是基于像素的，提取的精度普遍不高。提出了一种面向对象的道路震害信息提取方法，通过综合利用道路的多种影像特征及震前GIS矢量道路相结合来提取道路，然后依据提取道路的完整程度来识别道路震害信息。采用汶川灾区的遥感影像为例进行了实验，与目视判读的结果比较后证明该方法有效改善了信息提取的速度和精度。

关键词: 高分辨率遥感, 面向对象, 道路, 震害识别