Disaster assessment with parallel image processing for GIS based local area disaster decision support system

Abstract

Natural disasters, such as hurricanes, tsunamis and earthquakes cause huge loss of life, enormous amounts of property damage, and lengthy recovery times. Although it is impossible to avoid the costs of disasters, human sufferings can be minimized through effective disaster management and decision support system that can facilitate and expedite the resource distribution process more efficiently and effectively. This paper is to present the design of GIS based disaster data management, visualization and decision support system for North West Indiana region utilizing grid computing and visualization resources at Purdue University Calumet (PUC). One of the key factors in the disaster management system is to provide damage assessment maps in timely manner. We propose a parallel image processing algorithm utilizing computational grid to compute the disaster damage assessment map from pre- and post- disaster satellite images. Disaster decision support system will compute the effective resource distribution strategy and prioritize the rescue areas by utilizing disaster specific geospatial information system. This research will lay a foundation for disaster preparedness, management and decision support system at local government agency level.