Damage Detection Technique based on Texture Analysis

Remotely sensed data have been utilized efficiently for damage detection immediately after the natural disaster since they provide valuable information on land cover change due to spatial synchronization and multitemporal observation over large areas. Damage information obtained at an early stage is important for rapid emergency response and recovery works. Many useful techniques to analyze the characteristics of the pre- and post-event satellite images in large-scale damage detection have been successfully investigated for emergency management. Since high-resolution satellite images provide a wealth of information on damage occurred in urban areas, they are successfully utilized for damage detection in urban areas. In this research, a method to perform automated damage detection is proposed based on the differences of the textural characteristics in pre- and post- high resolution satellite images.