Signal De-Noising Based on Improved Matching Pursuit Algorithm

The Matching Pursuit (MP) sparse decomposition and reconstruction algorithm based on the over-complete dictionaries has been widely used in signal processing field, by which the signal structure interrelated atom can be extracted self-adaptively, so the noise suppression can be get. However, most of the computation in MP must used to find the best matching atoms, which will lead to many deficiencies such as slower processing speed and lower matching precision of the algorithm. In view of the above, in this paper, an improved Particle Swarm Optimization(PSO) algorithm with gradient information was put forward to find the best atom during the MP sparse decomposition process, which can get more efficiency and more accuracy of MP algorithm. Based on the MP algorithm optimized by improved PSO, signal decomposition reconstruction and de-noising Matlab simulation experiments were done, and simulation results show that the computation efficiency and signal reconstruction accuracy of the proposed algorithm in this paper is higher than the basic MP algorithm, and the signal de-noising effect is better than wavelet de-noising.