Extraction of the bare rock rate in counties of typical karst areas based on unmanned aerial vehicle images and satellite data, China
Bibliographic Data
| ID | 21649596 |
|---|---|
| Authors | Dongliang Wang (0000-0003-4388-7863, Key Laboratory of Land Surface Pattern and Simulation, Institute of Geographic Sciences and Natural Resources Research Chinese Academy of Sciences Beijing China, corresponding author), Xiangyu Sui (0000-0002-1885-1298, Key Laboratory of Land Surface Pattern and Simulation, Institute of Geographic Sciences and Natural Resources Research Chinese Academy of Sciences Beijing China) |
| Year | 2023 |
| Volume | 34 |
| Issue | 17 |
| Pages | 5499-5513 |
| Publication date | 2023-11-01 |
| Open Access | Yes |
| Type | ARTICLE |
| Venue | Land Degradation and Development (JOURNAL) |
| Journal identifiers | ISSN: 1085-3278 • E-ISSN: 1099-145X |
| Publisher | Wiley (PUBLISHER • GB) |
| DOI | 10.1002/ldr.4860 |
| OpenAlex | W4386093926 |
| Language | EN |
| References cited | 32 |
Rocky desertification is a land degradation process. The bare rock rate ( BRR ) is an important index. Satellite‐based methods for BRR estimation can generate BRR datasets, but the spatial resolution and accuracy are low. Unmanned aerial vehicle ( UAV ) data can be used for BRR extraction with much higher resolution and higher accuracy, but the data are laborious and costly to collect. This study proposed a stepwise upscaling method employing the UAV ‐extracted BRR as the ground truth to establish a multiband regression model and calibrated the BRR derived from 30‐m resolution LANDSAT 8 data, which was then used to calibrate 500‐m resolution moderate resolution imaging spectroradiometer ( MODIS ) data. We compared the results with the satellite‐based BRR and the results of UAV‐MODIS upscaling methods. The results showed that (1) the UAV provides high‐resolution data, and exposed rocks can be efficiently extracted from UAV images. This method could replace ground surveys. (2) The accuracy of the stepwise upscaling method was higher than that of the nonstepwise upscaling method in large areas. R 2 increased by 64.9%, and the root mean square error ( RMSE ) decreased by 82.9%. (3) Compared with that of the satellite‐based BRR , the accuracy of the multiband regression model‐extracted BRR was higher. R 2 increased by 21.5%, and the RMSE decreased by 62.2%. This study demonstrated that the stepwise upscaling method and multiband regression can be combined to extract high‐precision and large‐scale BRR data, which can be used to serve ecological engineering monitoring and ecological governance
Ground truth · Image resolution · Mean squared error · Remote sensing · Satellite · Statistics · Computer Science · Engineering · Environmental Science · Landslides and related hazards · Mathematics · Remote Sensing in Agriculture · Soil erosion and sediment transport · Artificial Intelligence · Geology
| Highly cited | No |
|---|