Integration of hurricane wind analysis and multimedia semantic content analysis for public outreach

Natural disasters, such as hurricanes, could have an enormous impact on society. The level of the public's preparedness could make a significant difference in the severity of casualty and damage inflicted by such storms. We present a prototype system to reach out to the public and improve their awareness of the potential dangers involved with such weather events. This Web-based system aggregates H*Wind storm track and wind fields data along with relevant videos extracted from YouTube and displays it to the user using Google Earth. A content-based concept detection algorithm is used to extract the videos, which may describe the impact of the storm in relevant geographic locations. Using Hurricane Ike as a case study, the result demonstrates how some of the information collected and displayed by the system could have increased the awareness of the public and potentially helped prepare them better to the devastating storm.