Design of Compressed Sensing Coding Matrix in Remote Sensing Imaging System

Lulu ZHANG,Quansen SUN,Jixin LIU

DOI: https://doi.org/10.3969/j.issn.1009-8518.2017.05.009

2017-01-01

Abstract:As a new way to break through the traditional Nyquist signal acquisition theorem, compressed sensing has received widespread attention in recent years, especial in remote sense imaging. The design of the observation matrix in compressed sensing plays a very important role. In fact, whether the original signal can be effectively compressed and then be accurately reconstructed in the receiving device, largely depends on the quality of the observation matrix. However, the observation matrix at present has a serious drawback: it is hard to be implemented; especially in remote sensing image sampling a simple and power-saving equipment is in-dispensable. The purpose of this study is to find a matrix with good performance and also easy to be imple-mented. In view of the advantages of the sub block compression perception in the reconstruction speed and quality, the two-level scale encoding matrix design was proposed based on the sub blocks, that is, the first block was divided, and then the encoding matrix was set. Through the simulation of optical mask sampling process, the design of the encoding matrix was analyzed and optimized and then the origin image was reconstructed with high quality in the terminal device.