Dataset · DannHiroaki
China building footprints geo-dataset mirror (CMAB)
DannHiroaki/China-Building-Footprints-CMAB-Mirror
A mirror of the scientific multi-attribute building footprints dataset covering all of China.
What's inside
A mirror of the CMAB geospatial dataset published on Figshare in 2025, described by the authors as the first national multi-attribute building dataset. Metadata indicates geospatial and text data in Chinese and English, though the Hugging Face mirror size is marked as under 1K records.
What you can do with it
- Train and validate building segmentation models on satellite imagery
- Geospatial analysis of urban development and building attributes
- Languages
- Chinese, English
- Size
- up to 1K rows
- Downloads this month
- 20.4K
- Likes
- 0
How to load it
from datasets import load_dataset
ds = load_dataset("DannHiroaki/China-Building-Footprints-CMAB-Mirror")