Detecting Cassava Plants under Different Field Conditions Using UAV-Based RGB Images and Deep Learning Models

A significant number of object detection models have been researched for use in plant detection. However, deployment and evaluation the real-time as well crop counting under varying real field conditions is lacking. In this work, two versions a state-of-the-art model—YOLOv5n YOLOv5s—were deployed evaluated cassava We compared performance when trained with different input image resolutions, images growth stages, weed interference, illumination conditions. The were on an NVIDIA Jetson AGX Orin embedded GPU order to observe models. Results case farm showed that YOLOv5s yielded best accuracy whereas YOLOv5n had inference speed detecting plants. allowed more precise counting, which mis-detected performed better interference at cost low speed. findings work may serve reference making choice model fits intended real-life application, taking into consideration need trade-off between speed, accuracy, memory usage.