Zuo Rui Shen, ipmist@cau.edu.cn, Ling Wang Gao, lwgao@cau.edu.cn, and Jian Wei Zhang, stiffant@yahoo.com.cn. China Agricultural University, Entomology, 2 Yuanmingyuan Xilu Road, Haidian District, Beijing, China
A national priority project has been conducted since 2001 for distance identification of crop pests, trying to solve such problem that the need for insects and plant disease identification is going up acceleratively. This problem becomes severe because of the scarcity of professional resources in insect identification. Demonstration sites of this project, located in Fujian province (east coast) and Chongqing province (southwest China), have shown a great progress. A web-based system, called as DidPests, has been developed and run over the Internet. It consists of three sub-systems. The first can make the identification automatically for over 40 species of insects with accuracy of 95% in minimum, by taking morphometrics, computer vision and artificial nerve-network (ANN) technology. The second is an expert system which can assist users interactively to identify insects and plant diseases, availing multimedia data of more than 230 insect species and 120 vegetable diseases in its database up to now. In fact, the final is an e-conference system, which provides a facility for experts and users to exchange their data and knowledge in the identification process. More species have to be identified automatically through training the system with more samples of insect imagery. More species have to be added into the multimedia database for expanding the range of assisted identification for both insects and plant diseases. And the DisPests system will be settled soon into the national center for monitoring and forecasting of crop pests, located in the Department of Agriculture, Beijing, so it will be able to serve for users over the nation.