Apprehension of Plant Diseases by Exploiting Machine Learning and Deep Learning Precepts

Plants play a pivotal role in the regulation of climate, agriculture, and economies; however, they are confronted with threats from diseases caused by bacteria, fungi, and viruses. The changing climate and weakened crop immunity have led to a surge in crop diseases, causing significant damage to crops and financial losses for farmers. Identifying and treating these diseases early is challenging due to the variety of diseases and limited farmer knowledge. Computer vision, combined with deep learning, offers a solution by leveraging leaf texture and visual features for disease identification. Plant health care involves anticipating and diagnosing life-threatening plant diseases, and early detection is key to reducing plant mortality. This study utilizes machine learning, a type of artificial intelligence, to build early prediction models for plant disease diagnosis. Machine learning addresses the difficulty of crop classification during early growth stages by employing drones with high-resolution optical imagery. These drones capture images of phonological stages, which are then used to develop characteristics based on grey level co-occurrence matrices. Various machine learning approaches, including random forest-nearest neighbors, linear regression, Naive Bayes, neural networks, and support vector machines, were employed to develop plant disease detection models. Evaluation metrics such as true positive rate, true negative rate, precision, recall, and F1-score were used to assess model performance. The results indicate that the ensemble plant disease model outperforms other proposed models. These predictive models aim to identify diseases in early stages, enabling timely preventive actions and predictive maintenance for plant health.