Deep Learning for Agricultural Plant Disease Detection

Abstract

Maize is the third significant crop in India. A number of fungal deceases are affecting Maize. The deceases Northern Blight, Southern Blight and Rust are causing reduction in the yield to an extent of 28–91%. The symptoms of these diseases appear on the leaves usually at the flowering stage. In this paper, study on automated detection of fungal disease in maize leaves is carried out. The proposed system uses image processing techniques using deep learning. The results obtained are encouraging with an accuracy of 98% for three class problems and accuracy of 70.5% for four class problems.