Multi-Crops Leaves Diseases Classification Using Fuzzy Logic and Pre-trained CNN Methods

Abstract

A nation's economy depends on the foundation of its agricultural sector, and the vitality of this sector is essential for the economy's steady growth. However, the frequency of plant leaf diseases threatens the quantity and quality of agricultural yields. It is important to classify these diseases effectively to identify and treat affected areas in plant images. Even though there are a lot of research publications on how to classify plant leaf diseases, automated detection models still need to be made more accurate; in this research, two crops widely grown in agricultural fields worldwide—cotton and tomatoes—will be further classified. Our approach entails creating a novel framework focusing on Artificial Intelligence (AI) techniques, particularly Deep Learning (DL)  methods. Fuzzy logic edge detection rules were adopted to generate reliable datasets. After that, we employ deep learning models including AlexNet and GoogleNet, for classification purposes. In our experiments, we build datasets with varied parameter tuning values, using 70/30 or 80/20 ratios for training and testing. The results show that the AlexNet method works better than others. It is 99% accurate, and the error rate is as low as 0.01%. These findings highlight the DL Methods revolutionize the classification of plant leaf diseases. This presents an opportunity for enhancing agricultural outcomes and ensuring economic stability.