2026-10-12T12:21:55Z https://www.scidb.cn/oai

10.57760/sciencedb.ecodb.00334 2026-08-12T09:42:10Z

2026-08-12 A Multi-Temporal UAV Multispectral Dataset for Land Cover Classification in a Heterogeneous Agricultural Landscape doi:10.57760/sciencedb.ecodb.00334 en This dataset presents a multi-temporal unmanned aerial vehicle (UAV) remote sensing dataset covering a typical farmland–wetland composite landscape in Jurong City, Zhenjiang, Jiangsu Province, China. The dataset comprises seven monthly observation periods from June to December 2025, each including RGB orthophotos, four-band multispectral imagery (green, red, red edge, near-infrared) and a digital surface model (DSM). Field surveys and remote sensing interpretation, conducted in parallel with the UAV aerial photography, provided ground-truth labels for 17 land cover classes, covering rice, dryland crops (maize, groundnuts, sweet potatoes), melons and vegetables (watermelons, loofahs, small vegetable plots), woodland, scrubland, nurseries, natural grassland, cultivated grassland, reed wetlands, ponds and reservoirs, residential areas, roads and greenhouse facilities. The seven-period time-series data provide comprehensive and in-depth coverage of the phenological cycle during the growing season, supporting the development and validation of phenology-aware classification algorithms. It is suitable for research into deep learning-based semantic segmentation, time-series classification and the fusion of spectral and structural data. UAV; multispectral; land cover classification Shangxiao Wang Shengjun Xiao Yanwei Sun Xiaonan Niu Zong Leli Liu Yi Zhang Ming EMBARGO https://creativecommons.org/licenses/by/4.0/ dataset Science Data Bank