A New Approach for Online Denoising
This paper discusses the shortcomings of conventional denoising methods, and introduces a new approach for online denoising. The low-pass and average filtering methods, the most favored and widespread online denoising techniques, have their inherent limitations caused by operational characteristics. These methods can eliminate the noise of high-frequency, but cannot reduce the noise of low-frequency without loss of the information of interest signal. To solve the problem of the simple denoising methods, wavelet analysis has been applied for the denoising, and the wavelet filter has significantly increased the performance of denoising because of some advantages over Fourier analysis. The wavelet filter, however, is unsuitable for online processing, which is the main drawback of wavelet analysis. To overcome the problems of the conventional denoising methods, this paper suggests a new approach which guarantees high performance denoising and can be easily actualized via online processing. The main concept of the suggested denoising method is to estimate the long-term signal of interest using a single long-term moving window and multiple short-term moving windows. The denoising method based on the estimation of long-term signal can reduce the noise of all frequencies without the loss of the information of interest signal. In addition, the suggested denoising method, which uses the moving windows, is highly suitable for online signal processing. The efficient performance of this suggested denoising method is confirmed by some experiments.
No Reference information available - sign in for access.
No Citation information available - sign in for access.
No Supplementary Data.
No Article Media
Document Type: Research Article
Publication date: 2012-05-01
More about this publication?
- The growing interest and activity in the field of sensor technologies requires a forum for rapid dissemination of important results: Sensor Letters is that forum. Sensor Letters offers scientists, engineers and medical experts timely, peer-reviewed research on sensor science and technology of the highest quality. Sensor Letters publish original rapid communications, full papers and timely state-of-the-art reviews encompassing the fundamental and applied research on sensor science and technology in all fields of science, engineering, and medicine. Highest priority will be given to short communications reporting important new scientific and technological findings.
- Editorial Board
- Information for Authors
- Subscribe to this Title
- Terms & Conditions
- Ingenta Connect is not responsible for the content or availability of external websites