The Research On Information Processing In High Resolution Microwave Staring Correlated Imaging

In disaster monitoring,resource exploration and other fields,high-resolution staring imaging technology is very important.For the traditional real aperture radar imaging technology,the resolution is limited by the antenna aperture,we have to construct a very large antenna,thus this technology is very difficult to apply.Microwave staring correlated imaging which is a new radar staring imaging technology proposed in recent years,can break through the resolution constraints of the antenna aperture,with the potential for high-resolution imaging.This thesis which is focused on the relevant key technologies of information processing can effectively promote the application of microwave staring correlated imaging technology in the field of remote sensing.The key technologies covered in this thesis include the following aspects.Firstly from the perspective of information theory,the information model of microwave staring correlated imaging is established and the quantitative characterization of imaging performance is given out.Further study on how to design the stochastic radiation source using the information theory tools to improve the capability to access information is given.The dissertation also gives construction methods of high-performance imaging equation.On this basis,the solving methods of imaging equation are studied,the processing algorithms to highly ill-conditioned imaging equations,the multi-grid and non-uniform grid approach for large scenarios and the echo processing methods based on narrow pulse stochastic radiation source are proposed.The detailed descriptions are as follows.In the first part of the thesis,the construction of information model and the design of radiation source based on information theory are carried out.Since the information theory was first proposed by Shannon in 1948,it is widely used in communication,radar and other fields.This thesis constructs the information model on the basis of mathematical model to describe the imaging process from the perspective of information theory,and proposes the transfer entropy and coupling entropy of imaging system,the mathematical expressions are derived to quantitatively characterize the ability of accessing information of imaging system.Under the information model,using the tools of information theory,the method to design radiation source is proposed based on temporal-spatial distribution entropy,the optimization method combined with genetic algorithm is carried out to design the parameters of radiation source,the design of the radiation source can maximize the use of limited aperture resources,bandwidth resources,realize the optimize configuration of the radiation source parameters.For the application on the space distribution of antenna elements,the element distribution entropy is proposed to achieve quantitative characterization of the randomness of element arrangement,and the optimization method using this entropy as the optimization goal is studied,the simulation results show the effectiveness of this method.The construction methods of microwave staring correlated imaging equation are studied in the second part of thesis.Imaging equation construction is a key step of the information processing of imaging,the imaging equation connects the radiation field distribution and the echo,by solving the imaging equation,we can get the imaging results of the imaging area.Traditional method is to construct the imaging equation by echo sample directly,it is very sensitive of the noise and non-ideal factors.This section presents construction methods of imaging equation based on time-domain integration,the simulation results demonstrate the effectiveness of this method.Further more,a variant pulse width time-domain integral equation construction method which enhances the randomness of the observation matrix by transmitting variable pulse width signal is proposed,we also give out a imaging equation construction method based on differential operation,the simulation results prove the effectiveness of the methods.A series of processing methods of highly ill-conditioned imaging equation are proposed in the third part of thesis.By the analysis of ill-conditioned problem of imaging equation,the characterization of ill-posedness based on the orthogonal degree of effective matrix column vectors is given.Furthermore several imformation processing methods are proposed,such as landweber algorithm,krylov subspace algorithms.Compared to the early proposed regularization algorithm,these algorithms don't change the equation structure,ensure the image resolution in condition of low SNR.Compared to the compressed sensing algorithm,these algorithms can get high resolution imaging result of complex scenes without prior sparse conditions.The correlation algorithms for the big image scene problem are proposed based on multi-grid and non-uniform grid methods in the fourth part of the dissertation.Firstly,the processing approach based on cascadic multi-grid(CMG)method is studied.The CMG method is a kind of preconditioned technology of solving a matrix equation,the original equation is decomposed into a set of coarser scales,both the fast-change error and the slow-change error are eliminated,thus we can solve the large-scale imaging equation effectively.The thesis also gives the non-uniform grid approach based on the distribution of target,through the non-uniform grid methods,we can realize large area imaging using less grid number.In order to get the imaging result quickly,the echo processing method based on narrow pulse random radiation source is proposed in the fifth part of the thesis.By emitting narrow pulse signal,we utilize the high range resolution of narrow pulse signal to shorten the time of processing.Combined with the correlated processing methods,we can realize high resolution image.Firstly the mathematical model of narrow pulse radiation source is established,and the echo processing flow is given.Then,the presision method to determine the start point of echo based on the weighted frequency domain spectral entropy is proposed.Finally,the weighted tape splicing method is studied.