M. Ferrara and G. Arnold (USA)
SAR, motion estimation, autofocus, registration, near-field invariants, invariant theory
Synthetic aperture radar (SAR) has proven to be a valu able imaging tool due to its all-weather, day or night ca pability. Typical SAR systems employ tomography-based reconstruction algorithms which depend on a known flight path of the sensor (specifically, the relative motion between the radar and the scene is needed in order to backproject the data onto an image surface). Typically, GPS technol ogy is used to provide position estimates along the sen sor’s flight trajectory. Unfortunately, GPS satellites may may not provide sufficiently accurate position estimates, or worse, may not be available in many defense and civilian reconnaissance applications. In order to eliminate depen dence upon GPS-based position estimates, we propose a scheme, based on invariance theory, which estimates the three-dimensional platform motion relative to persistent scatterers on the ground with absolutely no assumptions on the platform’s flight path, knowledge of the scatterer loca tions, or antenna beampattern.
Important Links:
Go Back
