Bam Earthquake Damage Mapping from Satellite Imagery

Damage mapping of the 2003 Bam earthquake by correlating pre- and post-event QuickBird satellite imagery, using pixel-based and object-based methods.

Introduction

Remote sensing is an effective science and tool within the disaster-management cycle; one of its applications is post-earthquake damage assessment and damage-map production. With high-quality satellite imagery, monitoring of the affected area can be accelerated, and given pre- and post-event imagery together with statistical analysis and image processing, the desired result is reached in the shortest possible time.

Pre- and post-event satellite imagery

To assess damage and map the changes in the city of Bam, pre- and post-event satellite images were examined. These images were acquired from the QuickBird satellite and are of high quality.

Pixel-based method

In the pixel-based method, the correlation between pre- and post-event images is computed with a moving window. It is simple and fast but of limited accuracy, and pixel information leaking from neighbouring features introduces error, which is partly mitigated by choosing an appropriate window size.

Object-based method

The object-based method removes the error caused by information leaking from neighbouring features and achieves higher accuracy. Based on the damage map and observations, it provides a sound estimate of the actual damage, and the change map was drawn at three damage levels: low, moderate and severe.