A hybrid regularized azimuth super-resolution imaging method

Abstract

The invention discloses a mixed regularization azimuth super-resolution imaging method, which is applied in the field of radar imaging and aims at traditional L 2 Regularized imaging method has low resolution, traditional L 1 For the problem of poor anti-noise performance of the regularized imaging method, the present invention first establishes a convolution-like model of scanning radar azimuth echo; then, through L 2 The regularization method preprocesses the echo signal; finally, based on the preprocessing result, L 1 The norm characterizes the target sparse prior information to reconstruct the target function, and the method of iteratively reweighting the norm is used to solve the target function. The innovation of the present invention lies in: combining L 1 Regular term and L 2 The advantages of the regular term, use L 2 The norm weakens the influence of noise and avoids the noise being amplified in the iterative process, and then uses L 1 The sparse nature of the norm improves the azimuth resolution of imaging. This method is different from the conventional L 2 Compared with the regularization method, the azimuth resolution is improved, and compared with the traditional L 1 Compared with the regularization method, the noise is well suppressed.

Description

technical field [0001] The invention belongs to the field of radar imaging, in particular to a forward-looking imaging technology suitable for airborne scanning radar. Background technique [0002] Airborne scanning radar forward-looking imaging has a wide range of applications in battlefield reconnaissance, material airdrops, aircraft autonomous landing, and precision guidance. Due to the small change of Doppler gradient in the forward-looking imaging area and the problem of angle ambiguity, traditional SAR and DBS methods cannot achieve forward-looking imaging. [0003] The working mode of the real-aperture scanning radar is relatively simple, and the high-resolution technology in the range direction is mature, but the resolution in the azimuth direction is still limited by the radar system. Since the azimuth echo can be regarded as the convolution model of the target scattering coefficient and the antenna model function, the deconvolution processing of the echo signal ca...

AI Technical Summary

Problems solved by technology

This method can suppress noise by smoothing the result, but the resulting resolution is limited

In the paper "Zou Jianwu, Zhu Mingbo, Li Wei, and Dong Wei." L for radar azimuth super-resolution 1 Norm regularization and its constraint method. In "Electro-optic and Control 22, no.8(2015): 33-36.", a method based on L 1 Norm-regularized azimuth super-resolution imaging method, which obtains high resolution in azimuth by adding sparse priors, but the ill-conditioned nature of the convolution matrix will limit the imaging performance at low signal-to-noise ratios

Images