The new large-scale satellite imagery dataset CHN6-CUG of representative cities in China selected six cities with different urbanization levels, city scales, development degrees, urban structures and historical cultures, including Chaoyang District in Beijing, Yangpu District in Shanghai, Wuhan city center, Nanshan District in Shenzhen, Shatin in Hong Kong and Macao in China. According to the degree of pavement coverage, the labeled roads include both track-covered and non-track-covered pavements. Based on the physical perspective of geographic factors, the marked roads include railroads, highways, urban roads and rural roads.
CHN6-CUG contains 4511 labeled images of size 512×512, divided into 3608 for model training and 903 for testing as well as result evaluation with a resolution of 50 cm/pixel. The original image of the dataset is in .jpg format and the road labeling image is in .png format, and the compressed data size is 175 MB.
The dataset is manually labeled pixel-level high-resolution satellite imagery with remote sensing image basemaps from Google Earth.
Based on pavement coverage, marked roads include track-covered roads and non-track-covered roads. Based on the physical perspective of geographical factors, marked roads include railroads, highways, urban roads and rural roads.
The data quality and accuracy are relatively high.
This work is licensed under
CC BY 4.0 (Creative Commons Attribution 4.0 International License).
CSTR:11738.11.NCDC.NIEER.DB6677.2024
10.12072/ncdc.nieer.db6677.2024
