A Real-Time Video Surveillance-Based Framework for Early Plant Disease Detection Using Jetson TX1 and Novel LeafNet-104 Algorithm

The precise and timely identification of plant diseases plays a pivotal role in safeguarding crop health and optimizing agricultural yields. This paper introduces an innovative framework designed for plant disease detection using video surveillance-based cameras strategically deployed on the Nvidia Jetson TX1 hardware platform within agricultural farms. Leveraging the capabilities of deep learning and the novel neural network architecture known as LeafNet-104, this framework empowers real-time disease detection on plants. Through the deployment of video surveillance cameras, continuous monitoring and early disease detection become achievable, enabling farmers to take prompt corrective measures and mitigate further crop damage. The Nvidia Jetson TX1, with its robust high-performance computing capabilities, enables on-device processing, rendering it ideal for real-time applications. To assess the efficacy of the proposed framework, comprehensive experiments were conducted on an actual agricultural farm. The dataset encompassed various plant species and disease types. Evaluation metrics such as accuracy, precision, recall, and the F1-score were employed to gauge the performance of the LeafNet-104 algorithm. The experimental findings underscored the framework's capacity to accurately and swiftly detect and classify plant diseases in real time, attaining an impressive accuracy rate of 97.5%.