ON THE USE OF THE IDEAL FILTER CONCEPT FOR IMPROVING SAR IMAGE QUALITY

The conventional matched filter forms the basis for SAR image generati on. Although the impulse response function obtained by the matched fil ter approach has the best achievable signal-to-noise ratio, high sidel obes, paired echoes and azimuth ambiguities will appear in the impulse response function (IRF) for severe phase and amplitude errors. This p aper considers two types of error: 1) SAR transceiver chain errors and instabilities which influence the range IRF and 2) azimuth signal ali asing which leads to azimuth ambiguities. In order to suppress these e rrors and to improve the SAR image quality, an ideal filter concept is used which provides, in addition to the matched filtering, the deconv olution of the phase and amplitude errors. Several simulation results are presented concerning the improvement of the range and azimuth IRFs . In the case of range compression, the sidelobes can efficiently be s uppressed if the replica of the transmitted signal is known with suffi cient accuracy. For azimuth processing, the Doppler parameters' accura cy and the azimuth fading limits the efficiency of ambiguity suppressi on. The results of the image quality analysis show that the sidelobes of the IRF obtained by matched filtering are attenuated by more than 1 0 dB when using the ideal filter concept.