EGU24-2353, updated on 08 Mar 2024
https://doi.org/10.5194/egusphere-egu24-2353
EGU General Assembly 2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.
Oral | Monday, 15 Apr, 11:35–11:45 (CEST)
 
Room 1.61/62

Automated glacier extraction using a Transformer based deep learning approach from multi-sensor remote sensing imagery

Yanfei Peng1, Jiang He1, Qiangqiang Yuan1,2, Shouxing Wang1, Xinde Chu3, and Liangpei Zhang4
Yanfei Peng et al.
  • 1Wuhan University, School of Geodesy, China(qqyuan@sgg.whu.edu.cn)
  • 2Hubei Luojia Laboratory, Wuhan, Hubei 430079, China(qqyuan@sgg.whu.edu.cn)
  • 3College of Geography and Environmental Science, Northwest Normal University, Lanzhou 730070, China
  • 4State Key Laboratory of Information Engineering, Survey Mapping and Remote Sensing, Wuhan University, Wuhan, Hubei 430079, China

Glaciers serve as sensitive indicators of climate change, making accurate glacier boundary delineation crucial for understanding their response to environmental and local factors. However, traditional semi-automatic remote sensing methods for glacier extraction lack precision and fail to fully leverage multi-source data. In this study, we propose a Transformer-based deep learning approach to address these limitations. Our method employs a U-Net architecture with a Local-Global Transformer (LGT) encoder and multiple Local-Global CNN Blocks (LGCB) in the decoder. The model design aims to integrate both global and local information. Training data for the model were generated using Sentinel-1 Synthetic Aperture Radar (SAR) data, Sentinel-2 multispectral data, High Mountain Asia (HMA) Digital Elevation Model (DEM), and Shuttle Radar Topography Mission(SRTM) DEM. The ground truth was obtained for a glaciated area of 1498.06 km2 in the Qilian mountains using classic band ratio and manual delineation based on 2 m resolution GaoFen (GF) imagery. A series of experiments including the comparison between different models, model modules and data combinations were conducted to evaluate the model accuracy. The best overall accuracy achieved was 0.972. Additionally, our findings highlight the significant contribution of Sentinel-2 data to glacier extraction.

How to cite: Peng, Y., He, J., Yuan, Q., Wang, S., Chu, X., and Zhang, L.: Automated glacier extraction using a Transformer based deep learning approach from multi-sensor remote sensing imagery, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2353, https://doi.org/10.5194/egusphere-egu24-2353, 2024.