Improving the accuracy of weather satellite images, using World Data Bank II geographic location data

Source citation:

• Deng, Z., & Ma, X. (2026). A frequency-based approach to improve the geometric accuracy of FY4B/AGRI geostationary satellite observations. Science of Remote Sensing, 13.

Deng and Ma evaluated a year’s worth of data captured with the Chinese FengYun-4B (FY4B) meteorological satellite, one of the new generation of geostationary satellites that ring the globe. Deng and Ma identified significant alignment errors it was making in the images it captured. The two then developed an automated system to fix those errors, enabling the capture of more accurate images critical for monitoring natural disasters and tracking environmental changes. Their complex system in part requires comparing textures and patterns in the satellite images to known geographic features, in order to calculate exactly how much the images need to be shifted to match reality. They used a detailed and highly processed geographic database to serve as a reliable reference for coastlines and inland lakes. The database combines boundary mapping coordinates from two datasets in the public domain: the World Vector Shorelines and the World Data Bank II. The raw dataset, World Data Bank II: North America, South America, Europe, Africa, Asia, created by the CIA to enable digitized mapmaking, is available in the ICPSR Member Archive.

Posted on March 26, 2026