Development of an image based expert system for identification of rice diseases and their management

摘要

Rice is an important cereal crop of Punjab state which produces 11 per cent and 2 per cent of the total rice production in India and world, respectively. The production of rice is widely affected by various rice diseases of which bacterial leaf blight(Xanthomonas oryzae pv. oryzae), sheath blight (Rhizoctonia solani), false smut (Ustilaginoidea virens), brown leaf spot (Drechslera oryzae), foot rot {Fusarium moniliforme), blast (Pyricularia grisea) etc affect the rice yield both quantitatively and qualitatively in the Punjab state. Diagnosis of diseases at early stages enables us to apply the control measures well in time and help in avoiding the yield losses. This process requires an expertise to identify the disease and prescribe the treatment. In present study, a computerized programme on image based rice disease expert system is developed to help extension personals, researchers and farmers in identification and management of these diseases. This expert system is bilingual and is developed inboth English and Punjabi language, and uses a hierarchical classification and step by step descriptive and visual methods of plant's disease diagnosis. The system involves four main sub tasks, namely, selection of language, plant part infected/affected,diagnosis and management of disease. The relevant content and recommendations were scouted from Punjab Agricultural University, Ludhiana and validated the same with the help of concerned experts. The system is designed and developed using Visual Basic 6as front-end and Microsoft Access-2003 as back-end software.