Extraction Building Damage for 2011 Tohoku, Japan Earthquake Tsunami Using Remote Sensing Images

Rapid assessing disaster degree and the disaster information spatial distribution is the key to achieve effective rescue after the geological disaster, and the building damage information extraction provide the basis for assessment. In this paper, used the object-oriented classification technology to extract building damage information of the pre- and post-earthquake tsunami in World-view2 image of sendai Japan in March 2011. Firstly, classify the feature threshold of different image objects and establish classification rules based on the multi-scale segmentation. Then used the classification rules on the optimal segmentation scale to classify and extract information for two images. And evaluate the accuracy and change detection for the results of classification. Finally, the vector data of change analysis are imported into the ArcGIS software by transfer matrix to calculate the building collapse rate. The results show that the building collapsed rate is 82.14%. It can provide reference data for the affected degree of evaluation.