A great number of disasters have been occurred around the world every year and cause extensive damage in various perspectives. When a disaster happens, it is crucial to determine its region, causes and severity of damage in order to improve the effectiveness of the required response. In this project, we aim to determine the severity of damage and its causes with the help of high-resolution aerial images via deep learning methods fed by pre and post images of affected regions. The images were taken from the xBD dataset. The xBD dataset is a recently published, one of the largest high-resolution satellite imagery datasets that contain pre-disaster and post-disaster images around the world for six different types of natural disasters.
