Research Of Three-dimensional Imaging Algorithm Based On Sliding-spotlighr SAR

Using high-precision SAR image of surface features target three-dimensional imaging is one the hottest topics in recent years, we use sliding spotlight imaging patterns to improve image accuracy azimuth traditional methods. We use compressed sensing algorithm to solve the problem that baselines are sparse and uneven. Traditional methods of three-dimensional imaging have steep demend of baseline arrangement. Using the new method to improve the usefulness of tomography has important practical significance.Focusing on the title ‘Three-dimensional Imaging based on Slinding-Spotlight SAR’, we study the Slinding-Spotlight SAR, image registration, conventional tomography and new tomography based on compressed sensing algorithm to solve the problem that the baselines are sparse and uneven. These problems have been made a deep study in different chapters. The principles of each part of the algorithm have been analysed, solutions have been gaven. Based on the study of the above problems, we analyse the weaknesses of the current approach, put forward some innovative algorithm. Generally speaking, the work of the paper are presented as follows:1 Sliding Spotlight SAR imaging problem, according to the sliding spotlight SAR imaging model, analyze Sliding Spotlight SAR echo signal, imaging accuracy, Doppler course. According to these knowledges, we do derivations sliding spotlight SAR imaging algorithm derived from striped, spotlight SAR imaging algorithm.2 image registration problems, because of the exist of baseline, the images of different aircrafts of different angles changes in some way. When the baseline is long, and the Terrain scene fluctuate obviously, the offset will lose regularity, pixel offset is no longer a fixed constant. The offsets of different position are different. So we propose a method based on image segmentation to solve such problems.3 Problem of height space focusing, height dimension focusing is the core issue of SAR tomography. We focuses on the aspects in this article, learn traditional tomography principle, grasp its resolution precision, imaging constraints and the impact of various factors on the imaging results.4 traditional tomography algorithms, including FFT algorithm and APES spectral estimation algorithm, has stringent requirements of the baselines. The aircrafts must be evenly distributed and the length of the baselinse must be shorter than the requirement of sampling theorem of Kuinaisite. But it often costs too much to meet these requirements in real life. We use an algorithm named compressed sensing to solve this problem that we can’t get effective images by using FFT or APES algorithm in the situation of sparse and uneven baselines. The core idea of compressed sensing sparse is to reconstruct the signal with under-sampling. It’s very suitable for the situation of sparse and uneven baselines. In this paper, we use two compressed sensing algorithms, OMP algorithm and COSAMP algorithm and verify the validity in the situation of sparse and uneven baselines. After further study of the compressed sensing algorithms, we prove that COSAMP algorithm has higher imaging precision than OMP algorithm.