A review of results obtained at the Institute of Atmospheric Optics SB RAS on passive sensing of the surface and atmosphere characteristics based on topical processing of satellite images in visible and infrared spectral ranges is presented. The first stage of the investigation was concerned with estimation of disturbing influence of atmosphere on the underlying surface temperature, which resulted in development of a self-radiating model for the atmosphere-ocean system in 3-5 and 8-13 μm transparency windows. Most important factors affecting the radiation attenuation in these ranges have been studied. Using the model, the accuracy of absolute calibration of the thermal channel (10.4-12.6 μm) of the space-borne MSU-SK instrument (Kosmos-1689 and Kosmos-1939) was performed. Further, main regularities in the image formation in visible and infrared spectral ranges when observing through inhomogeneous multicomponent scattering and absorbing atmosphere were studied; the available methods of atmospheric correction of the satellite measurements were mastered and new ones were elaborated. In the last decades, new efficient algorithms of satellite data topical decoding for various applications were worked out, e.g., on-line monitoring, express-detection of forest fires from space, and so on.