author(s)
H Ravisankar, K Sivaraju, D Damodar Reddy and N Srinivasa Rao

abstract
Tobacco quality and production are widely affected by diseases in both nursery and field crop. The yield losses due to diseases can be as high as 50% depending on the type and intensity of disease(s). Primary diseases affecting the tobacco are frog-eye spot, brown spot, black shank, damping-off, wilt, root-knot, mosaic and leaf curl diseases. An accurate and rapid diagnosis can help to take up appropriate and timely corrective measures and thus minimize yield losses. However, the end users are not able to utilize the available information developed by scientists. Symptoms of some diseases may sometimes be similar to nutritional disorders and hence confusing. Information technology created new vistas of scientific information, social interaction and communication. Based on information technology, Web based expert system on various diseases of tobacco, symptoms and their management was developed and linked to “Agridaksh” managed by IASRI for providing global access. The diseases on tobacco were classified into minor and major in both nursery and field crop and the database was designed in tree structure. The system has two different methods for disease diagnosis. The first method uses rule based inference (information system) and the second method uses ontology. The former is text based while the latter also shows pictures with the text. In both the approaches the user enters into the question-answer session with the system and at the end of the session, a particular tobacco disease is identified and its images get displayed online. Under this system, a knowledge model has been created in such a way that includes a knowledge acquisition, ontology based problem identification and a knowledge retrieval system for tobacco diseases. This system is useful to enhance the capacity of farmers, agricultural extension personnel and tobacco crop development agencies for tobacco disease management to increase crop yield. The expert system developed is of great utility as it brings disease symptoms and damage on tobacco and their management to a single platform.