Monitoring the distribution of clouds and snow cover is essential for climate, hydrological, and natural hazard management studies. This paper presents a Python-based algorithm for the automatic extraction and mapping of clouds and snow using satellite data acquired by the MERSI sensor onboard the FY-3D platform. The methodology relies on processing visible, near-infrared, SWIR, and thermal infrared spectral channels, combined with the computation of spectral indices such as the Normalized Difference Snow Index (NDSI). The algorithm applies combined spectral thresholds (red and NIR reflectance, brightness temperature, and NDSI values) to separate the target classes: clouds (white) and snow (blue). The results are exported as colored rasters (GeoTIFF) and vectorized into interoperable formats (GeoJSON, KML) for GIS applications and visualization in Google Earth. The implementation integrates spatial interpolation of geolocation data, polygon simplification, and automatic legend generation. Tests carried out on FY-3D datasets have demonstrated the method’s effectiveness in distinguishing between cloud-covered and snow-covered areas, with strong potential for extension toward near real-time operational monitoring. The proposed approach combines algorithmic simplicity with the flexibility of integration into geospatial processing workflows, providing a practical tool for climate, agricultural, and water resource management applications. The findings highlight the relevance of using Chinese FY-3D sensors for regional and global monitoring, as well as the importance of developing open, reproducible, and adaptable algorithms for diverse satellite data sources.

remote sensing, FY-3D MERSI, NDSI, cloud-snow detection

geodesy engineering

Presentation: poster