Research And Implementation Of Emergency Cross-Media Information Retrieval System

In recent years, more and more emergencies like natural disasters and public health events happens, and they threaten the health and security of the people seriously. So the acquisition and monitor of emergencies information is needed. At the same time, image is becoming an important information carrier, the identification of image is becoming more important. So this thesis researched and implemented the emergency cross-media information retrieval system. The thesis had completed the following tasks:(1)Designed a method of simulating user login to collect the image information of emergency. Added an algorithm of URL filtering based on contents correlation evaluation into the Heritrix framework to implement the collection of multimedia information.(2)Created an algorithm of image semantic analysis based on feature map and done the semantic analysis by getting the visual features of image. Used the query expansion technology based on ontology to realize the query expansion, and added the text retrieval of Lucene to realize the retrieval of multimedia information.(3)Designed and implement the intelligent mobile terminal of the system. In the design, the system used the stratified architecture and created a method of condition monitoring to develop the intelligent mobile terminal.(4)Used the agent communication mechanism based on KQML and the agent coordination mechanism based on management agent to implement the agent manager of the system, optimize the allocation of resources and develop the efficiency of the system.In this thesis, the emergency cross-media information retrieval system provided the retrieval service of emergency multimedia information on the PC machine and the intelligent mobile terminal, realized the acquisition and monitor of emergencies information and protected the people’s health and security. Besides, system had a certain theoretical significance and application value.