Winter UAV Remote Sensing Dataset for Farmland Boundary Detection: A High-Resolution Multi-Terrain Agricultural Image Collection

This study introduces a high-resolution UAV (Unmanned Aerial Vehicle) remote sensing image dataset aimed at advancing the development of deep learning-based farmland boundary extraction techniques and supporting the optimal deployment of Solar Insect Lights (SILs). Agricultural pests pose a significant threat to crop health and yield, while traditional pest control methods often cause environmental pollution. Solar Insect Lights offer an environmentally friendly and efficient alternative, with their deployment needing to be customized according to the unique characteristics of each farmland. Field ridges, as natural boundaries, are ideal locations for placing Solar Insect Lights. However, accurately extracting ridge information from remote sensing images is crucial for achieving this goal. This dataset, captured using UAV technology, includes high-resolution images of various farmland types with different topographical and environmental conditions. These images, with sub-decimeter GSD (Ground Sample Distance), provide rich detail for deep learning models to support more accurate farmland boundary extraction. The release of this dataset aims to address the scarcity of data in this field.