AI RESEARCH
Beyond GSD-as-Token: Continuous Scale Conditioning for Remote Sensing VLMs
arXiv CS.CV
•
ArXi:2605.07562v1 Announce Type: new Remote sensing vision-language models (RS-VLMs) face a fundamental mismatch with natural-image counterparts: the same geographic object exhibits radically different visual evidence across ground sampling distances (GSDs) spanning multiple orders of magnitude. Yet existing RS-VLMs often discard GSD or inject it as a discrete text token, forcing a single static parameter set to absorb the entire scale spectrum. We