Abstract: After earthquake, it's a key for relief to access timely to road damage information and analyse traffic capacity. However, traditional method of using remote sensing technology to extract damage information has the problems of low efficiency, and affects the timeliness of earthquake emergency. In our daily training, we focus on a class of sensor images and get the most appropriate parameters, and combining the object-oriented classification, we propose a specific process including preprocessing, image classification, rapid road damage identification. Experimental results show that the method can improve obviously the extraction rate in keeping with a good of classification precision. So it is expected that the method will be applied with much practical uses.

Key words: High resolution remote sensing image, Object-oriented classification, Road damage identification