Feature Extraction for Diseased Leaf Image Classification using Machine Learning

Recognition algorithms for crop disease are based on the extraction from diseased plant leaf images of different types of features. Leaf diseases are important factors as they can lead to a significant reduction in agricultural crop quality and quantity. Therefore, detecting and understanding diseases is an important task. The approach to leaf image-based disease recognition consists of two steps: I extracting color and shape characteristics from lesion images; (ii) classifying diseased leaf images using machine learning approaches. This paper analyzes the efficiency of the classification performed using Support Vector Machine, K-Nearest Neighbor and Decision trees based on the extracted characteristics.