Application Research Of Agricultural Expert System Based On Embedded Platform

With the rapid development of artificial intelligence technology, application ofmodern agricultural information technology in agriculture represented by agricultureexpert system is more and more widely. At present most of the agricultural expertsystems are PC-based development, although this type of agricultural expert systemscan help farmers solve agricultural problems, considering that it is inconvenient tocarry and facilitate the direct guidance of agricultural production in the front line, so itlimits the kind of agriculture expert system spread. In order to solve the problem ofagricultural expert system’s inconvenient to carry,this paper studied agricultural-basedportable embedded expert system, combining with the requirements of today’sagricultural development.Firstly，this paper studied the construction of the ARM11-based embedded systemhardware and software platform, and introduced development of embedded modelsand processes in detail，as well as how to build embedded development environment,including: the establishment of cross-compiler environment, Bootloader transplant,Linux kernel porting and root production of documents.Secondly, the agricultural expert system knowledge representation, inferencemechanism and database design were studied in this paper. Based on a diagnosis ofplant pests in agricultural expert system, this paper designed to produce aformula-based knowledge representation, inference mechanism to forward reasoningas the inference engine according to the characteristics of the portable agriculturalexpert system, combining database and knowledge on the inference in order to becloser to the data, thus it improved the efficiency of reasoning and queries.In addition, almost all agricultural expert systems are used method directly entertext or select the input method in human-computer interaction. Although this does notaffect the performance of the agricultural expert system, this paper proposes a newhuman-computer interaction taking into account the application of the field conditionsand the operation of portable devices, namely image input method, as the traditionalinput method supplement. Users can take pictures for a particular plant pests anddiseases by a portable device camera, then system will analyze photographs extractfeature information to extract feature information as the basis for agricultural expertsystem inference to draw conclusions or recommendations. The contents of the imagefeature extraction is complex, and there are follow-up specialized research topics, the input scheme design and camera driver has been studied in this article for the future toachieve this input basis.Finally, this paper designed the agricultural expert system software architecture,and studied portable agricultural expert system. It has contributed to the wideapplication for modern agricultural information technology in agriculture.