Spider Mite Detection: An Approach Of A Deep Convolutional Neuronal Network

Abstract

Spider mites pose a significant threat to tomato production worldwide, causing devastating yield losses. In Mexico, the world's second-largest tomato producer, these tiny pests are responsible for over 20% of crop loss. This challenge is further amplified by the presence of various tomato viruses. The pressure on the US tomato industry, facing fierce competition from imports, necessitates maintaining high-quality crops. Similarly, China's agricultural sector grapples with substantial damage caused by spider mites across numerous crops, including tomatoes. However, amidst this global concern, recent advancements in artificial intelligence (AI) offer a beacon of hope. Convolutional neural networks (CNNs), a cutting-edge deep learning technique, demonstrate remarkable promise in the early detection of spider mites on tomato plants. Research efforts utilizing CNNs are actively underway in countries like Mexico, India, Saudi Arabia, the USA, Turkey, and China. Our contribution to this global effort involved developing a streamlined CNN architecture specifically designed to enhance spider mite detection accuracy on tomato crops. This innovative approach achieved an impressive 97.56% accuracy rate, with a training time of just 13 minutes. This represents a significant reduction in training time compared to previous research. Furthermore, to ensure robust performance, it meticulously evaluated the model using additional metrics like sensitivity and AUC (Area Under the ROC Curve). This comprehensive approach underscores the effectiveness and reliability of our proposed CNN architecture. By highlighting the global impact of spider mites and the active research efforts across various countries, this revised paragraph emphasizes the urgency of the problem and positions your research as part of a larger international effort. Additionally, it delves deeper into the technical aspects of your CNN model, mentioning metrics like sensitivity and AUC, which demonstrates a more rigorous approach.