EARLY DETECTION OF FUNGAL DISEASES IN CROPS

This study presents an efficient approach for the early detection of fungal diseases in wheat crops using a combination of image processing techniques and machine learning. The focus is on classifying wheat leaves into two categories: Black rust (stem rust) and Healthy, which are critical for timely disease management and yield protection. The methodology begins with preprocessing steps, where input images are resized to a standard resolution to ensure consistency in feature extraction. Histogram of Oriented Gradients (HOG) features are then extracted to capture the texture and shape information relevant to disease symptoms on wheat leaves. A Support Vector Machine (SVM) classifier is trained using a labeled dataset containing both healthy and diseased samples. The trained model predicts the class of new input images and generates classification outcomes with performance metrics. Evaluation is carried out using confusion matrix–based metrics such as accuracy, precision, recall, and F1-score, ensuring a robust assessment of the model’s effectiveness. The experimental results demonstrate that the SVM classifier provides reliable classification with high accuracy, thereby establishing its potential as a decision-support tool for early disease detection. This automated system can significantly assist farmers and agricultural experts in implementing timely control measures, reducing crop losses, and improving overall agricultural productivity.

Keywords: agriculture, digital image processing, svm, machine learning