Anthracnose Disease Detection in Cashew Leaf Using Machine Learning Technique Based on Contour Detection and Principal Component Analysis

Abstract

Detecting and classifying leaf diseases in cashew crops is critical for farmers to find pest and disease infections. Cashew leaf diseases can reduce productivity if not detected early. Creating an automated method utilizing image processing for leaf disease identification decreases time and expense and primarily contributes to a rise in cashew nut yield. For image segmentation, canny edge detection and an active contour model are utilized. A feature extraction method, Principal Component Analysis (PCA), is applied when the contour has been applied. After the features have been extracted, they are submitted for categorization. This study analyzed several classifiers’ accuracy, precision, and recall values. These classifiers included Random Forest, SVM, KNN, and Naive Bayes. This research tries to answer whether a machine learning classifier provides the best results when the diseased area is divided using the canny edge detection and contour detection technique