LuoJia-FG

LuoJia-FG is a large-scale, attribute-driven, and hierarchical fine-grained land-cover dataset designed to bridge the "semantic gap" in high-resolution Earth observation. Unlike existing datasets with flat taxonomies and coarse categories, LuoJia-FG introduces a deep semantic structure tailored for next-generation Vision-Language Models (VLMs).The dataset comprises 119,619 multimodal triplets (Image, Mask, Text) covering 1.59 million km^2 across diverse biogeographical zones in China, with spatial resolutions ranging from 0.5m to 2.0m.