Full resolution focusing of Seasat SAR images in the frequency-wave number domain

Abstract

Synthetic aperture radar (SAR) data focusing has been traditionally performed using matched filter techniques. However, downward continuation techniques can produce, with a great computational efficiency, results that only the most sophisticated conventional techniques can achieve. The basic idea is to decompose the measured field into monochromatic plane waves of appropriate amplitudes, wave lengths and propagation directions. The backpropagation of these plane waves, and their recombination at the time when the sounding pulse was emitted, can produce a map of the backscattered electromagnetic field. In order to obtain correct focusing of synthetic aperture radar raw data, both the geometrical and the transmission parameters of the system should be known as precisely as possible. The transmission parameters are generally known very precisely, whereas the geometrical ones (i.e. sensor-target relative position, satellite velocity and attitude, etc.) can be derived from the ephemerides of the satellite or from the data themselves. The only two geometrical parameters involved in the focusing process are the sensor-target relative velocity and the sensor-target closest approach distance. In this paper we sugggest a new technique to extract these two geometrical parameters from the data themselves (autofocusing) in order to achieve the best possible focusing of SAR images. This technique has been tested both on synthetic and on real Seasat data. The achieved precision of the parameters value has been estimated to be better than one part in 10000.