An Aerial Weed Detection System for Green Onion Crops Using the You Only Look Once (YOLOv3) Deep Learning Algorithm

Addie Ira Borja PARICO, Tofael AHAMED
2020 Engineering in Agriculture Environment and Food  
The real-time object detection system You Only Look Once (specifically YOLOv3) has recently shown remarkable speed, making it potentially suitable for Unmanned Aerial Vehicle (UAV) precision spraying. In this study, YOLO-WEED, a weed detection system based on YOLOv3, was developed. The dataset, derived from a five-minute UAV video, was split into a 69 : 17 : 13 ratio for training, validation, and testing, respectively. YOLO-WEED demonstrated a real-time detection speed (up to 24.4 FPS) and high
more » ... performance using NVIDIA GeForce GTX 1060, with a mean average precision of 93.81 % and an F1 score of 0.94. These results successfully show the effectiveness of the YOLO-WEED system for real-time UAV weed detection, given its high speed and high accuracy in detection. [Keywords] you only look once (YOLO), deep learning, real time weed detection, convolutional neural network (CNN), unmanned aerial vehicle
doi:10.37221/eaef.13.2_42 fatcat:ugvw26nq7bd2hhtv5i3yrlunle