Satellite observing systems are producing image observations of the Earth's surface and atmosphere with spectral and spatial resolutions that result in data rates that current general-purpose computing systems are incapable of processing and analysing. As a result, current processing systems have been able to analyse only limited amounts of image data with less than optimal algorithms for generating high-quality geophysical parameters. A massively parallel processor (MPP) is operationally available at NASA/GSFC for routine image-analysis applications. Research studies with the MPP are being pursued in the area of interactive spatial contextual classifications for the land thematic mapper data, automatic SIR-B stereo terrain mapping, ice-motion detection, faint-object image restoration and other general purpose ocean and land image-processing systems. Several applications are presented comparing the MPP products with enhancements of imaging data with standard image-processing methods. Finally, a work-station parallel processor for space station on-board image processing will be described.