Leaf disease severity classification with explainable artificial intelligence using transformer networks

Abstract

The agricultural sector is considered India's most crucial sector. In India, farming is the most common occupation and a significant source of revenue [1]. India is second in the world in terms of population, with 70% of its people living in villages and relying primarily on agriculture for their subsistence. Farmers cultivate a wide variety of crops based on numerous factors such as crop conditions, the environment, soil conditions, local farming practices, new variants of pathogens and various illnesses, etc. *Author for correspondence Presently farmers are incurring losses due to changes in climatic conditions, plant disease is one of the main challenges to food security since it significantly lowers crop quality and output [2].Therefore, it has been difficult to diagnose diseases correctly and discover them early. A variety of methods, including soil management, crop rotation, irrigation, genetic modification, harvesting methods, precision farming, weed control, and pest and disease control, are used to boost the yield. So, it is crucial to identify pesticides and diseases early if you want to boost a crop's production. The traditional method of identifying plant diseases is by visual inspection [3]. This can result in incorrect disease identification, and