Geometric Correction and Accuracy Verification of the Geology-1 Hyperspectral Satellite
Geology-1 is China’s first hyperspectral remote sensing satellite designed specifically for geological and mineral industry applications, which was successfully launched into orbit on May 17, 2025. This article presents a systematic description of the geometric processing methods employed in the generation of Geology-1 Level-2 geometric data products, encompassing an on-orbit geometric calibration strategy based on a rigorous imaging model and the orthorectification product generation pipeline, along with a comprehensive evaluation of the geometric positioning accuracy under ground-control-point-free conditions. Experimental results demonstrate that the mean plane root-mean-square error (RMSE) of geometric positioning for the detectors is approximately 0.9 pixels, indicating that subpixel geometric positioning accuracy has been achieved. The relative band-to-band registration RMSE across the three detectors falls consistently within the range of 0.2–0.4 pixels. In multiorbit image overlap regions, the plane registration RMSE is better than 0.5 pixels, and adjacent-orbit orthorectified images maintain strong geometric consistency at orbital track boundaries. These results demonstrate that the proposed geometric processing framework achieves stable positioning performance across a variety of terrain conditions and is well suited for operational deployment, thereby providing a reliable geometric foundation for regional-scale applications of Geology-1 hyperspectral data and multiorbit mosaic product generation.