Automated Identification of Plant Bacterial Diseases Using Computer Vision Techniques

Abstract

Plant diseases pose a significant threat to global food security by reducing crop yield and quality. Accurate and early identification of such diseases is very important for effective management and prevention. Traditional disease identification methods rely on laboratory and visual inspection, which are time-consuming, labour-intensive, require expert knowledge and are often inaccurate. In recent years, computer vision and artificial intelligence (AI) have emerged as promising tools for automatically detecting plant diseases. This study presents an automated system for detecting plant bacterial diseases in leaf images via computer vision and deep learning. The datasets contain 1200 images representing healthy and infected leaves from crops such as tomatoes, rice and lemon, which were analyzed. Firstly, pest-associated symptoms were carefully screened and excluded through field level observations, integrating entomological knowledge to distinguish insect damage from bacterial infections. In addition, laboratory based bacterial isolation and biochemical characterization was performed to confirm presence of casual pathogens, ensuring the authenticity of the dataset. Image pre-processing techniques such as segmentation, resizing and normalization were applied to enhance the feature extraction. Convolutional Neural Networks (CNN) were used to classify bacterial diseases such as bacterial leaf blight, bacterial spot and citrus canker. The model achieved an overall accuracy of 93.4% with precision and recall values exceeding 90% for most disease categories. The results demonstrate that vision-computer-based systems can effectively differentiate between healthy and diseased plants based on visual symptoms such as discolorations, lesions and texture changes. This study highlights the potential of AI-driven approaches in plant pathology, offering rapid, cost-effective and scalable solutions for disease detection at the initial stage. The integration of such technologies into agricultural practices can improve disease management, be cost-effective and reduce reliance on chemical treatments. Future research should focus on improving model generalization and real-time field deployment for sustainable agriculture.