Disaster management using image processing

Abstract

Many hazards that threaten the country have the ability to cause loss of lives or injury and all of them have the ability to cause severe damage to homes, businesses and infrastructure. These includes earthquakes, meteorological hazards, accidental hazards and flooding. Disasters have become a severe issue of growing concern throughout the world, whether it is natural hazards or by human factors. Our country is vulnerable to a number of natural hazards. It is highly imperative to develop effective methods for disaster management. I propose and demonstrate an image processing technique to identify shortest possible route to affected area after disaster like earthquake, flood, Fire & volcano hazards. The proposed model composed of pre- processing, decision making and result. In my thesis I used speeded up robust features to segment roads based on pre and post disastrous moment. I also used color based segmentation to detect fire and volcano on roads and it also can detect floods on road. Moreover, I used a method called ‘K’ means clustering which detect the presence and absence of an object by comparing both pre & post disaster images. Finally, I used a shortest path estimation algorithm to find the best possible route to affected area so that the immediate relief can reach the site as soon as possible.