Resumo

CEDENO POZO, Antonio  e  TRUJILLO CODORNIU, Rafael. Online wavelet-based scheme for reduction of noise in industrial signals. Rev cuba cienc informat [online]. 2014, vol.8, n.3, pp. 12-26. ISSN 2227-1899.

This paper proposes a scheme for online Gaussian noise reduction based on wavelets with sliding windows. The scheme has been specifically designed for real-time systems (e.g. SCADA systems) where signals must be processed on-the-fly. The proposed algorithm differs from previous proposals on: i) the adaptive threshold selection, which is based on NeighShrink; ii) the dynamic use of previously calculated wavelet coefficients and; iii) the use of the Haar wavelet filtering implementation instead of others with higher computational cost. The novel scheme has been tested on a set of pattern signals proposed by Donoho and Marron complemented with other representative measurements obtained from real processes. Our results indicate that, for this kind of data, the proposed algorithm outperforms previous schemes on both: noise reduction and efficiency.

Palavras-chave : industrial signals; moving windows; noise reduction; wavelet transform.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )