Reconstruction algorithm based on compressed sensing

ZHANG Tao1，ZHONG Shun-cong1，2，3*，ZHU Zhi-bin4，5，FU Xi-bin5

（1. Laboratory of Non-destructive Testing & Evaluation，School of Mechanical Engineering and Automation，

Fuzhou University，Fuzhou 350108，China；2. Fujian Key Laboratory of Medical Instrument and Pharmaceutical Technology，Fuzhou 350002，China；3. Key Laboratory of Safety Science of Pressurized System of Ministry of Education，East China University of Science and Technology，Shanghai 200237，China；4. School of Chemical Engineering，Fuzhou University，Fuzhou 350108，China；5. Xiamen Special Equipment Inspection Institute，Xiamen 361000，China）

Abstract：Aiming at introducing and using compressive sensing（CS）into practice，CS was discussed in detail from sparse representation，encoding measurement and reconstruction algorithms. The commonly used reconstruction algorithms，such as matching pursuit，orthogonal matching pursuit and the regularized orthogonal matching pursuit，were reviewed. The minimum mean square error（MMSE） linear estimatealgorithm，which showed better reconstruction quality under low sampling rate，was introduced as well. Through the image

simulation，compared with commonly used reconstruction algorithms，the results indicate that MMSE shows great superiority and potential for real applications.