AUTOMATED DETECTION AND CLASSIFICATION OF PLANT LEAF DISEASES Using Digital Image Processing and Machine Learning

Article Information

Abstract

Plant diseases are a major cause of reduced agricultural productivity and frequently appear as visible changes in leaf colour, texture, shape and surface structure. Manual diagnosis depends on expertise and may be slow when large agricultural areas require monitoring. This paper presents a computer-aided framework for automated plant leaf disease detection using digital image processing and machine learning. The proposed pipeline consists of image acquisition, preprocessing, leaf segmentation, diseased-region identification, feature extraction, classification and performance evaluation. Colour statistics, Gray-Level Co-occurrence Matrix (GLCM) texture measures and morphological descriptors are considered as handcrafted features, while convolutional neural networks and transfer learning provide an alternative end-to-end approach. The paper discusses Support Vector Machine, Random Forest, K-Nearest Neighbour and CNN-based classification, together with accuracy, precision, recall and F1-score. Particular attention is given to dataset quality, field-condition variability, class imbalance and the risk of reporting inflated results from controlled images. The study concludes that a hybrid, carefully validated image-processing system can provide an efficient foundation for low-cost plant disease screening, particularly when combined with mobile and edge-computing technologies.

Keywords

DOI Active Record

This scholarly article has been successfully registered and assigned a persistent DOI through DOIGLOBAL.

Research Tools

Share This Research

Recommended Citation