An Enhanced Machine Learning Approach For Identifying Paddy Crop Blast Disease Management Using Fuzzy Logic

Abstract

Disease is a major issue in agriculture, particularly for rice crops, as it can significantly reduce yield and harm the health of the plants. In order to address this problem, various image processing and soft computing technologies have been developed to detect and mitigate disease in agricultural crops. However, in some cases, the detection of disease remains a challenge. In this research, a new automated method for detecting paddy disease is proposed by means of an optimized fuzzy interference system (OFIS). The projected method involves dividing collected paddy images into different colour bands and lowering noise in the greenish band using a median filter. The texture and color properties of the pre-processed green band are then removed and input into the OFIS scheme, which classifies the picture as heathy or sick. The OFIS system is a rule-based system that uses linguistic factors to classify data, and its parameters are optimized using the different step size firefly method. The performance of the projected system is evaluated in relationships of accurateness, sensitives, and specificity.