Remote sensing---the observation of Earth from instruments beyond it---turned geography's ancient problem, knowing the whole surface at once, into an engineering discipline, and its archive into the planet's longest continuous self-portrait. This article presents a narrative review of that arc's canonical line: Anderson and colleagues' 1976 land-use classification, Lillesand and Kiefer's 1979 textbook synthesis, Tucker's 1979 NDVI, Townshend and Justice's 1986 African vegetation dynamics, Turner and colleagues' 2003 biodiversity sensing, Woodcock and colleagues' 2008 free-Landsat advocacy, Wulder and colleagues' 2008 continuity analysis, Drusch and colleagues' 2012 Sentinel-2 mission, Hansen and colleagues' 2013 global forest change, Gorelick and colleagues' 2017 Google Earth Engine, Zhu and colleagues' 2019 free-data benefits, and Campbell and Wynne's 2011 educational synthesis. The synthesis is organized around three themes: observation, in which sensors, spectral signatures, and indices made the surface measurable; application, in which land use, vegetation, and biodiversity became monitoring programs; and openness, in which free data and cloud computing turned planetary analysis into an everyday act. It is concluded that remote sensing is geography's empirical revolution---the first instrument to see the whole Earth continuously---and that its history pivots on a policy decision, the 2008 Landsat opening, more than on any sensor. This record was produced with AI assistance (GLM large language model) as part of an automated writing project. It has not undergone peer review.