A Novel High Precision SAR Focussing Algorithm Based On Chirp Scaling

Abstract

Azimuth compression of synthetic aperture radar data is an inherently two-dimensional problem because the SAR reference function migrates through several range resolution cells. This effect is referred to as 'range cell migration'. For perfect focussing of the image the fact has to be accounted for that the amount of range cell migration varies with range. In classical range-Doppler SAR processors as well as with two-dimensional frequency domain processors a space-variant interpolation is required in order to compensate for this effect. In general, interpolation will degrade the final image quality and is computation time consuming. The proposed new algorithm avoids any interpolation step. It is essential for the algorithm that the SAR data are transformed to the range-Doppler domain prior to range compression. In this domain each range line is premultiplied with a particular phase function which is designed to perfectly straighten the range cell migration trajectories even for an arbitrarily wide swath by the subsequent range compression step. This new method for range migration correction is the key element of a new SAR processing algorithm described in detail. The proposed method is inherently phase preserving. Its focussing quality is neither limited by high squint nor by wide swath. The implementation is simple because only multiplications and Fourier transforms are required. It is essential for the new algorithm that the transmitted pulse is a linearly frequency modulated chirp of the form exp{jn*k*z2}.