Research On Compressive Sensing In ISAR Imaging

Inverse synthetic aperture?ISAR?has been widely employed in military and civil fields,because ISAR imaging has large role in distance and has little influence by the natural environment for moving target imaging.ISAR mainly observes moving target,and most of them are non-cooperative.So it is hard to guarantee enough observation time for every goal.In real life,some of ISAR echo may be lost in the spreading when they encounter noise or interference.Traditional range Doppler method is based on the complete ISAR data,so it can't guarantee good imaging for the missing data echo.A new signal processing theory Compressed Sensing?CS?can compress signal,at the same time signal sampling.So it can use few measurements to recover original signal,which is far less than the dimension of the original signal.This characteristic not only solves the traditional ISAR imaging problems of large data,big calculation and data redundancy,but also ensures the high quality imaging for the missing data echo.This article firstly introduces the imaging principle of ISAR,which includes the ideal model of turntable and motion compensation technology.And motion compensation technology is used in ISAR practical measured data.Then CS radar imaging technology is studied in this paper.This part research focuses on the basic theory of compressed sensing and the signal reconstruction methods.In order to verify the effectiveness of compressed sensing reconstruction algorithms,some experiments are done in the simulation sine signal and field plane practical measured data.Lastly,CS reconstruction algorithm is improved from two sides respectively.One is log-sum norm,which is equivalent norm to l₀ norm.Some theories of this article illustrate that the solving method of log-sum is easy and it is more approximation norm to l₀ norm than l₁ norm.The experiments show that log-sum norm not only has better signal sparse representation than l₁ norm,but also is stronger in anti-noise performance.The other is Gini Index,which has better sparse performance than l₁ norm and some other norms.So based on Gini Index CS reconstruction algorithm is proposed in this paper.More experiments are done in simulation sine signal and ISAR experimental measured data to verify the effectiveness of Gini Index.At the same time,some experiments are done to compare the performance of Gini Index with log-sum norm in many ways.