研究目的
To present first results of a snow cover detection algorithm applied to high resolution Sentinel-1 images of two different mountainous protected areas and compare them with snow maps generated from Sentinel-2 acquisitions.
研究成果
The implemented algorithm showed good performance in the flatter Hardangervidda area but had lower quality results in the more complex Sierra Nevada National Park. Future improvements include better dry snow classification and adaptive thresholding strategies.
研究不足
The algorithm has issues with snow cover detection in areas characterized by complex topography, climate, and land cover. Misclassifications occur due to backscattering changes not related to wet snow and the difficulty in distinguishing dry snow from bare soil areas.
1：Experimental Design and Method Selection:
A change detection based approach tailored to handle Sentinel-1 SAR intensity images was implemented. The algorithm is based on the methodology proposed in [5], tailored to handle Sentinel-1 SAR intensity images.
2：Sample Selection and Data Sources:
Sentinel-1 and Sentinel-2 images of Hardangervidda National Park (Norway) and Sierra Nevada National Park (Spain) were used.
3：List of Experimental Equipment and Materials:
Sentinel-1 and Sentinel-2 satellites, ASTER DEM, temperature maps.
4：Experimental Procedures and Operational Workflow:
SAR datasets were calibrated and terrain-corrected using an ancillary DEM. A 2D median filter was applied to mitigate speckle. The ratio between the backscattering coefficient of each winter image and the reference acquisition was calculated.
5：Data Analysis Methods:
The snow cover maps generated from Sentinel-1 were compared with those from Sentinel-2 to evaluate the algorithm's performance.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
