Zelong Wang, 国防科技大学


Sparse Modeling in Reconnaissance and Imaging


Electronic reconnaissance and imaging are important methods to obtain information, where the quality of signal and image processing is of great importance. However, it is difficult to obtain the proper solutions to many inverse problems in signal and image processing for their ill-posedness. To make them well-posed, the sparseness, one of the most important regularizations, is usually introduced. In this talk, we mainly study three information processing problems, i.e. DOA estimation, image processing, and image reconstruction, from the viewpoint of sparseness. Firstly, the backgrounds of these problems are introduced and the sparse priors of them are explored. Secondly, the models are established, such as signal subspace model for DOA estimation, regularized sparse representation model for image processing, and sparse high-resolution reconstruction model for both optical and microwave imaging. Lastly, we give some possible works in the future: more accurate models, faster algorithms, and practical system design.

« Back