One-dimensional Signal Denoizing Method Based on Segment Similarity

To improve the effectiveness of existing denoizing algorithms, we propose a definition of segment similarity in one-dimensional signals and a signal denoizing method.First, the denoized signal is obtained and divided into a certain number of signal segments by using an existing method into a certain number of signal segments.Then, in accordance with the defined segment similarity, we construct a set of similar segment collections for each signal segment and calculate the similar segment weights of the elements in the collection based on their similarity.Third, the signal segment of denoizing result is acquired based on the weighted average of all elements.All the denoizing results of the signal segments are combined to obtain the denoizing result of the entire signal.Finally, two experiments are conducted.Experiment results showed that compared with wavelets, principal component analysis, and sparse representation, the proposed method can improve the signal-to-noise ratio and reduce the mean square error effectively.The results verify the effectiveness and the feasibility of the proposed method.