Intelligent Pattern Recognition System with Application to Cotton Leaf Disease Identification

Abstract

The cultivation of cotton is a major source of cash for farmers in most of the regions of India. The production of cotton is affected due to leaf diseases. This paper illustrates technical know-how which identifies the leaf inflicted with the disease, and it segregates them according to the exact class of the disease. Through state-of-the-art image processing the image of inflicted leaf is captured in such a way that its background is kept intact using Otsu’s segmentation. The color, texture, and shape are cultured and fed to neural network for assimilation. The targeted population size of the disease taken is of three types, namely Bacterial leaf blight, Myrothecium, and Alternaria. The sample size is collected from CICR Nagpur and actual fields from Wardha and Buldhana districts. The precision obtained for categorization is 95.48%.