SAR moving target imaging method based on improved minimum entropy algorithm

Abstract

The invention provides an SAR moving target imaging method based on an improved minimum entropy algorithm. The method comprises the following steps: 1, deducing a two-dimensional frequency domain signal of an SAR moving target echo of a front side view, constructing a phase compensation function, extracting ROI data of a moving target after imaging by using a traditional range migration algorithm, and constructing a phase compensation function of the ROI data of the moving target; and 2, searching the optimal value of an unknown parameter in a moving target signal based on an improved minimization image entropy method, and performing phase compensation on the ROI data of the moving target to complete focusing of the moving target. The improved minimum entropy algorithm introduces the thought of the dichotomy into the iterative minimum entropy algorithm, the parameter value range of the next iteration is determined through the positive and negative of the derivative of the entropy function, the parameter value range is continuously narrowed, and finally the optimal parameter enabling the entropy function value to be minimum is obtained. The improved minimum entropy algorithm can overcome some defects of a traditional minimum entropy algorithm, and has the advantages of being high in convergence speed and high in adaptability.

Description

technical field [0001] The invention relates to signal and information processing technology, in particular to a SAR moving target imaging method based on an improved minimum entropy algorithm. Background technique [0002] Synthetic Aperture Radar (SAR) is an active, all-time, all-weather, high-resolution earth observation radar system that is widely used. Ground Moving Target Imaging (GMTIm) is an important application in SAR. The movement of ground moving targets will lead to Doppler shift, range unit migration and azimuth modulation frequency changes, and the traditional SAR imaging algorithm suitable for ground stationary targets is difficult to image ground moving targets. [0003] The existing ground moving target imaging methods can be divided into optimization methods, time-frequency transformation, Region of Interest (ROI) and so on. The time-frequency transformation type of moving target imaging method is based on Keystone transform (see Wan jun's "Refocusing of...

AI Technical Summary

Problems solved by technology

The iteration step size of the first method is constant, the calculation is inflexible and needs to traverse more times; the iteration step size of the second method is the absolute value of the gradient multiplied by a certain multiple, and the multiple is generally unchanged, and there are also The problem of inflexibility. In addition, the traditional minimum entropy method still has a certain hyperparameter sensitivity problem.

Images