Decision-level fusion of satellite imagery and LiDAR data for post-earthquake damage map generation in Haiti

Earthquake is one of the most lethal natural disasters and a severe threat to human life. The damage maps are precious information to mitigate the casualties after an earthquake in guiding the rescuers towards the affected area. This paper proposes a novel method, named decision-level damage estimation (DLDE), to generate the building damage map using post-event high-resolution satellite imagery (HRSI) and light detection and ranging (LiDAR) raster data. The meaningful information which describes the available data is produced through texture analysis in the primary step of the proposed method. Support vector machine (SVM) classification algorithm is employed to extract the damaged buildings after the separation of the building area. In the next step, the damage degree based on LiDAR and satellite image for each building is calculated, and in the final stage, these damage degrees are fused to obtain the final damage degree for buildings. We evaluated the proposed DLDE method using the WorldView II satellite image and LiDAR data of Port-au-Prince, Haiti, which was acquired after the 2010 earthquake. The overall accuracy (OA) of 81% proved the high ability of the proposed method for the assessment of post-earthquake damage.