This dataset is based on Landsat remote sensing images, and desert area images are obtained through pre-processing algorithms such as radiometric calibration and atmospheric correction. Sand ridges in desert areas are extracted through manual visual interpretation. The deserts involved in the dataset mainly include Taklimakan Desert, Gurbantunggut Desert (the Junggar Basin Desert), Badain Jaran Desert, Tengger Desert, Kumtag Desert, Qaidam Basin Desert, Kubuqi Desert, Ulanbuh Desert, including five phases of 2000, 2005, 2010, 2015, and 2020. This sand ridge dataset provides basic data for the extraction method of sand ridges in desert areas of China and the study of sand dune movement patterns.
Landsat land satellite imagery data: from the website of the United States Geological Survey（ https://www.usgs.gov/ ）Geospatial Data Cloud（ http://www.gscloud.cn/ ）Download.
1. Remote sensing data preprocessing: radiometric calibration, atmospheric correction, image stitching, cropping, color correction.
2. Manual annotation method for extracting sand ridges: Use edge annotation software tools based on PyQt5 to manually annotate sand ridges on cropped images, which become the training set for the model.
3. Data augmentation: Mirror flip and rotate the training set in 16 directions to obtain a 32 fold dataset.
The average accuracy of the annotated sand ridge line is within 1 pixel (30 m).
| # | number | name | type |
| 1 | 2022YFF0711702-04 | National key R & D plan |
This work is licensed under a
Creative
Commons Attribution 4.0 International License.
CSTR:11738.11.NCDC.NIEER.DB6508.2024
10.12072/ncdc.nieer.db6508.2024