YOLO-Seg Model with Tiling Movement for Detecting Paddy Parcels using High-Resolution Drone Image

Description

This study aims to develop a system that combines object detection and segmentation capabilities in high-resolution images using the YOLO-Seg model for paddy parcel monitoring based on drone imagery. For the first time, we propose a novel tiling movement technique and systematically analyze optimal overlap rates to reduce tile boundary errors that occur in large-scale drone image processing.

Categories

Licence