Snow Cover Detection Using Multi-Temporal Remotely Sensed Images of Fengyun-4A in Qinghai-Tibetan Plateau

Differentiating between snow and clouds presents a formidable challenge in the context of mapping cover over Qinghai–Tibetan Plateau (QTP). The frequent presence cloudy conditions severely complicates discrimination from satellite imagery. To accurately monitor spatiotemporal evolution cover, it is imperative to address these challenges enhance segmentation schemes employed for assessment. In this study, we devised pixel-wise classification algorithm based on Support Vector Machine (SVM) called 3-D Orientation Gradient (3-D OG), which captures variations gradient direction dimensions geostationary “Fengyun-4A” (FY-4A) multi-spectral multi-temporal optical This assumes that speed are different process movement leading their discrepancy characteristics time space. Therefore, algorithm, images calculated first, then angle trend obtained that. Finally, feature space composed image, maps, used as input SVM. Our results demonstrate proposed can identify more during snowfall by utilizing FY-4A’s high temporal resolution image. Weather station data, was collected snowstorms QTP, were evaluating accuracy our algorithm. It demonstrated overall using OG improved at least 12% 10% compared products Fengyun-2 MODIS, respectively. Overall, has overcome axial swing errors existing Geostationary satellites successfully applied cloud QTP. Furthermore, study underscores visible near-infrared bands Fengyun-4A be near real-time monitoring with performance