China-MAS-50k

We release China-MAS-50k (China Multi-label dataset for Agriculture & rural Scene 50k), the first very-high-resolution (VHR) remote sensing dataset for multi-label classification covering entire China’s agricultural and rural areas, filling the gap in finely annotated data for non-urban scene recognition. Based on a 50 km grid sampling system, over 50,000 sample points were determined nationwide, where VHR Google Earth imagery were to be collected for subsequent multi-label generation. A fine-grained label system comprising 18 categories (i.e., cropland, rural village, greenhouse, photovoltaic station, etc.) was established. Meanwhile, both a rigorously defined visual interpretation system and labeling procedures including cross-annotation, sample filtering and error correction were proposed to maintain annotation quality and reliability. Finally, the proposed dataset has a total of 55,520 VHR images with 135,289 labels. It exhibits a significant long-tail distribution, thereby providing a challenging benchmark for multi-label classification and long-tail learning.