In this work, the developed system for detecting areas with defects in the development of corn crops was investigated from a photograph taken by an unmanned aerial vehicle (UAV) using computer vision. To solve the problem of detecting such sites, the structures of the YOLOv5 and YOLOv8 neural networks were considered. The use of the developed software will reduce labor and time costs for image analysis, which in turn will reduce the response time when problem areas are detected in agricultural fields to achieve higher yields.
Keywords: instance segmentation, YOLOv5, YOLOv8