Studies And Establishment On Expert System For Identification And Chemical Control Of Paddy Weeds

In this paper the identification and control of 58 kinds of paddy weeds in south China was studied with the rule-based reasoning (RBR) and the case-based reasoning (CBR) technology, and an expert system for identification and control of paddy weeds (ICPWES) was established. When this system was tested with rule-based reasoning, the identified rate was above 67.8%, and if make use of images to identified weeds, it was found that the identified rate will be above 75%. The case-based reasoning module can provide proper control case to user according to the real occurrence of paddy weeds, if there are not a case equal to the input information by user, the system will provide a similar case according to the variety of weeds. The expert system own explaining function and can acquire knowledge by itself. In this system, the author have built a classify database of 58 kinds weeds that frequency occurrence in paddyfield, a database about those weeds' information, a herbicide database and a control case database. 290 weeds' images, a help file that help the user to use the system convenient and others files were attached in the system.Because kinds of computer technologies (multimedia et al) were used, this system would be provided with a friendlier man-computer interface, a stronger capability and easier operation; its knowledgebase and casebase could also be modified and enlarged.Result in this study showed that it was feasible completely to constructing weeds management expert system with rule-based reasoning and case-based reasoning technology. In the future the key of the study will be focus on how to improve the low identified rate and how to make the system more easy to be used.