One of the numerous fundamental tasks to perform rescue operations after the earthquake, check the status of buildings has been destroyed. The methods to obtain the damage map are two categories Shared. The first group of methods uses data before and after the earthquake, and the second group that only uses the data after the earthquakes that we want to offer a flexible and according to information that we are available to achieve the damage map. In this paper, we work on VHR satellite images of Haiti, and by UNet that is a convolution network. The learning algorithms profound changes to improve the results were intended to identify the damage of the buildings caused by the earthquake. The deep learning algorithms require very training data that it's one of the problems that we want to solve it. As well as Unlike previous studies by examining pixel by pixel degradation, ultimate precision to increase that show the success of this approach felt and has been able to reach the overall accuracy of 78.61%. The proposed method for other natural disasters such as rockets, explosions, tsunamis, and floods also destroyed buildings in urban areas is to be used.
1). Do you think that the proposed method is useful for other natural disasters such as rockets, explosions, tsunamis, and floods that also destroyed buildings in urban areas?
2). with VHR satellite images and UNet convolution network, is the overall accuracy of 68.71% of the earthquake damage assessment acceptable? .
2. This accuracy calculated from a large area. Many of previous studies calculate accuracy only in a small area which size of those buildings is normal, but in our case study about 50% of the buildings are very smaller than normal building like slum. All of the buildings inputed into the algorthim not only big buildings.
