Improved entropy-based autofocus correction for synthetic aperture radar

Abstract

Autofocus is an essential step in order to obtain the well-focused synthetic aperture radar image. Compared to the non-parametric autofocus, the parametric way is not subjected to the prominent targets so much. Entropy is widely used in parametric autofocus. It is sensitive to the clutter and not a convex function either. In this article, the authors present a new modified entropy to improve the robust of the parametric autofocus by reducing the number of the local optimal solutions remarkably. During the searching process, the residual azimuth phase error (RAPE) is decomposed as the summation of the polynomials. Of course, the used first derivative of the new entropy with respect to the model coefficients is given in detail. Moreover, for applying azimuth subaperture strategy, the authors remove the linear difference of adjacent RAPE via estimating the Doppler centroid, which is more effective than the map-drift method. Finally, real data experiment validates the authors’ approach.

Keywords

• radar imaging
• entropy
• synthetic aperture radar
• Doppler effect
• polynomials
• radar clutter
• local optimal solutions
• residual azimuth phase error
• synthetic aperture radar image
• parametric autofocus
• modified entropy
• searching process
• polynomials
• azimuth subaperture strategy
• Doppler centroid estimation
• improved entropy-based autofocus correction
• clutter