A Novel Citrus Canker Disease Detection Using GA And KNN Hybrid Model Based on Various Feature Descriptors

Abstract

With the advent of era of modernization the automation based systems are eliminating the need of human involvement. These automated systems are capable of achieving accuracies to highest possible level. This article proposes a image and machine intelligence based system for Canker disease occurring in Citrus fruits. The image of the fruit is enhanced in terms of contrast and a noise reduction filter is applied. After that the disease area is segmented to extract the local disease characteristics. The features extracted are selected by Genetic Algorithm based optimization. The reduced dataset is used to train Classifiers. The best results are achieved with GA and KNN hybrid approach.