HUGStereo: A Hierarchical Fusion and Geometry-Aware Refinement Network for High-Resolution Satellite Stereo Matching

End-to-end stereo matching methods designed for terrestrial imagery often perform poorly on satellite data due to complex scenes with abundant occlusions, repetitive and textureless regions, and large variations in scale. To address these challenges, we propose a novel network architecture, HUGStere...

Olles dieđut

Bibliográfalaš dieđut

Váldodahkkit:	Yuxiang Yao, Lina Yang, Haoyan Yang, Pengyuan Sun, Lianhui Liang
Materiálatiipa:	Artigo
Giella:	Inglês
Almmustuhtton:	IEEE 2026-01-01
Ráidu:	IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Fáttát:	Convolutional neural network disparity estimation hierarchical cost fusion satellite stereo images stereo matching
Liŋkkat:	https://ieeexplore.ieee.org/document/11503019/
Fáddágilkorat:	Lasit fáddágilkoriid Eai fáddágilkorat, Lasit vuosttaš fáddágilkora!