Approximate message passing-based SAR sparse feature enhanced imaging method

Abstract

The invention discloses an approximate message passing-based SAR sparse feature enhanced imaging method. The method comprises the steps of: establishing a Lasso imaging mathematic model for a synthetic aperture radar echo complex data; deducing a synthetic aperture radar image analysis sparse solution in the Lasso imaging mathematic model by utilizing an approximate message passing algorithm; andobtaining a sparse feature enhanced SAR image on the basis of the synthetic aperture radar image analysis sparse solution by utilizing a Gauss-Seidel method. The method has the advantages as follows:an AMP algorithm is utilized to solve the problems of Lasso optimization, and approximation is carried out through utilizing a central limit theorem and taylor expansion, so that the large calculatedamount caused by message passing can be effectively reduced. In addition, compared with the existing algorithm, the AMP algorithm is capable of greatly improving the convergence speed and making the image recovery quality good without increasing the operand of single-step iteration, so that the AMP algorithm is very suitable for the large measurement matrix-related compression sensing problems.

Description

technical field [0001] The invention belongs to the field of synthetic aperture radar (Synthetic Aperture Radar, referred to as SAR) imaging technology, in particular to a SAR sparse feature enhancement imaging method based on approximate message passing (Approximate Message Passing, referred to as AMP). Background technique [0002] Synthetic aperture radar belongs to microwave imaging radar. It can realize two-dimensional or three-dimensional high-resolution imaging of ground scenes by using limited aperture antenna. [0003] In order to obtain high-resolution SAR images, it is necessary to obtain and process broadband radar signals, but this also means that the radar data will increase greatly. In order to solve the large amount of radar data, people have introduced and applied compressed sensing technology. Compressed sensing technology means that as long as the signal is compressible or the signal is sparse in a certain transform domain, an observation matrix unrelated ...

AI Technical Summary

Problems solved by technology

Many methods have been proposed to improve the accuracy of radar imaging, such as the OMP algorithm. However, due to the coupling of range and azimuth in radar imaging, the observation matrix is ​​too large, and the imaging process will face a large amount of storage and calculation. Therefore, These compressive sensing radar imaging methods are difficult to implement, and then the IST (Iterative Shrinkage Thresholding) algorithm was produced. This algorithm is based on an approximate observation model, and the quality of radar imaging is better. Inspired by the decoupling operation of traditional imaging algorithms, the distance Solve the direction and azimuth separately, which can quickly and effectively reduce the calculation amount of single-step iteration of radar imaging, but because the convergence speed of the IST algorithm is very slow, and the overall calculation amount is also large, so it is urgent to solve this problem

[0005] In summary, how to improve the existing radar IST algorithm to speed up the convergence of the algorithm is also a new challenge

Images