STEREO RADARGRAMMETRY WITH DIGITAL SIDE-LOOKING RADAR IMAGES

摘要

With self-calibrating parameters in polynomials, basic radargrammetric conditions are extended to allow for residual systematic errors in side-looking radar (SLR) imagery. In a further attempt, we apply a statistical interpolation method, that of linear prediction with filtering, to describing time-varying orientation parameters. We document briefly this algorithmic development over the last 4 years at our Center. It is felt we implicitly take advantage of some experiences gathered in geometric anlysis on optical remote sensing imagery. Using stereoscopic real-aperture side-looking airborne radar images over a hilly area near Taipei, we conduct tests about the effects resulting from functional models extended for SLR point determination. While rather limited accuracy data are obtained, we learn that weight constraints on over hundreds of unknown parameters are very important to radargrammetric processing for positions.