Comparison of Phase-Gradient Autofocus and Sharpest-Image Autofocus in ISAR Imaging

The phase-gradient autofocus and the sharpest-image autofocus are two classical non-parametric algorithms for phase correction in inverse synthetic aperture radar (ISAR) imaging. This paper analyzes the principles of the two algorithms and compares the performance of the two algorithms. The sharpest-image autofocus estimates the correction phase to maximize the sharpness of the image. This method optimizes the image directly and thus has an excellent performance. The phase-gradient autofocus estimates the correction phase from some isolated strong scatterers. Its performance may be degraded when the strong scatterers cannot be well isolated from other scatterers. Therefore, the sharpest-image autofocus has a better performance than the phase-gradient autofocus, especially for distributed targets. The results of field data confirm our analysis.