Effective Cultivated Land Extraction in Complex Terrain Using High-Resolution Imagery and Deep Learning Method

Remote Sensing, Год журнала: 2025, Номер 17(5), С. 931 - 931

Опубликована: Март 6, 2025

The accurate extraction of cultivated land information is crucial for optimizing regional farmland layouts and enhancing food supply. To address the problem low accuracy in existing products poor applicability methods fragmented, small parcel agricultural landscapes complex terrain mapping, this study develops an advanced model western part Henan Province, China, utilizing Gaofen-2 (GF-2) imagery improved U-Net architecture to achieve a 1 m resolution mapping terrain. We obtained optimal input data by fusing spectral features vegetation index from remote sensing images. evaluated validated effectiveness proposed method multiple perspectives conducted change detection landscape fragmentation assessment area. experimental results show that achieved F1 score 89.55% entire area, with ranging 83.84% 90.44% hilly or transitional zones. Compared models solely rely on features, feature selection-based demonstrates superior performance adjacent mountainous regions, improvements 4.5% Intersection over Union (IoU). Cultivated parcels are smaller than 0.64 hectares. From 2017 2022, overall area decreased 15.26 km2, most significant reduction occurring areas, where fragmented. This trend highlights urgent need effective management strategies prevent further loss these areas. anticipate findings can contribute precision agriculture modernization terrains world.

Язык: Английский

BGSNet: A boundary-guided Siamese multitask network for semantic change detection from high-resolution remote sensing images

ISPRS Journal of Photogrammetry and Remote Sensing, Год журнала: 2025, Номер 225, С. 221 - 237

Опубликована: Май 5, 2025

Язык: Английский