Stepping up Tomato Leaf Diseases Detection using YOLOv10

Abstract

Early detection of diseases in tomato plants is a crucial factor in increasing agricultural productivity and reducing the risk of crop failure. Diseases that affect tomato leaves can significantly reduce yields if not detected early. Conventional methods that rely on manual observation by farmers are often inefficient, require specialized expertise, and have varying levels of accuracy. Therefore, this study implements the YOLOv10 (You Only Look Once version 10) method to automatically detect tomato leaf diseases through image analysis using deep learning techniques based on Convolutional Neural Networks (CNN). The objective of this study is to develop a YOLOv10-based tomato leaf disease detection system that enhances disease identification accuracy, accelerates early detection processes, and provides a practical and accessible technological solution for farmers. Furthermore, this research aims to optimize the application of computer vision technology in agriculture to improve decision-making efficiency and reduce reliance on conventional methods that are less effective.The results of the study indicate that the YOLOv10 method achieved a detection accuracy of 95.3%, with a precision of 94.8%, recall of 93.7%, and a mean average precision (mAP) of 95.6%. The developed application has an average inference time of 0.15 seconds per image, providing real-time detection results with low power consumption. The implementation of this technology has been proven to reduce pesticide use by up to 40%, increase farmers' decision-making efficiency by 70%, and help minimize the risk of widespread disease transmission. These results demonstrate that utilizing the YOLOv10 method for tomato disease detection can enhance disease identification accuracy, expedite early detection, and offer an innovative solution to support smart, efficient, modern, and sustainable agriculture.