International Journal of Science, Engineering and TechnologyAuthors: Sahana S A, Lisha N Singh, Priyanka R N, Rekha D
Abstract: Agriculture plays a vital role in India’s economy, and early detection of plant diseases is essential to prevent crop loss and ensure food security. Traditional manual inspection methods are often inefficient and prone to error. This paper presents a lightweight and interpretable plant disease detection system based on classical machine learning and image processing techniques. The proposed approach uses K-means clustering for image segmentation to isolate infected regions from plant leaves. Texture and colour features are extracted using the Grey- Level Co-occurrence Matrix (GLCM) and HSV colour histograms. These extracted features are classified using a Support Vector Machine (SVM) to identify various plant diseases. The system achieves high accuracy while remaining computationally efficient, making it suitable for low-resource environments. By avoiding complex deep learning architectures, the proposed model ensures faster processing and better interpretability, supporting early disease diagnosis and promoting sustainable agricultural practices.
DOI: https://doi.org/10.5281/zenodo.17512518
