Typhoon Disaster Information Extraction And Dynamic Visualization

China is one of the few countries that affected by the typhoon most severely. The typhoon often caused heavy losses to the southeast coastal people every summer and autumn. Because typhoon is the focus of disaster prevention and reduction in the coastal areas, so it is of great significance to obtain typhoon-related information rapidly, comprehensively and accurately.Today, the internet media, which can response rapidly to news and events, has been completely popular, and a large number of the typhoon disaster related information is existed in the form of electronic documents or news reports. To extract typhoon disasters information from these materials will be helpful to master the latest developments of the typhoon disaster, and provide relevant data to support decision making on disaster prevention and reduction rapidly. In the above context, this study took typhoon disaster-related news documents as the corpus to extract disasters information with the help of natural language processing techniques. Firstly, this study used ICTCLAS (Institute of Computing Technology, Chinese Lexical Analysis System) to segment the related documents. And because ICTCLAS was lack of capacity to identify place name and typhoon disasters field words segmentation, this study built domain dictionary of typhoon disasters and introduced place name dictionary to assist ICTCLAS in segmentation together. Then, this study used natural language processing technology to extract published time of news and disasters address, and extracted disasters time and disasters type information manually in the corpus by the human-computer interactive approach.This study developed geocoding function of typhoon disasters address based on the Google Maps API. However, place name/address, particularly administrative units, with strong currency resulted in place names in different periods indicated different spatial extent. Therefore, taking Xiamen city as example, address match of place name in different periods was implemented by time, place names/address information and the administrative division historical data of Xiamen city. The dynamic display of the typhoon disasters information was realized by combining with types of typhoon disaster according to the time order, based on disasters address matching. In the end, the above function was integrated into typhoon disasters information extraction function and dynamic display system which was developed based on GeoDatabase and ArcGIS Engine9.2. This System completed the extraction of typhoon disasters information in the news document corpus, and also provided an intuitive and imaginal displayed way for typhoon disasters information.