A Digital Impulse Noise Removal Filter

Autor: Kathy J. Dien Hillig, Michael D. Morris
Rok vydání: 1982
Předmět:
Zdroj: Applied Spectroscopy. 36:700-701
ISSN: 1943-3530
0003-7028
DOI: 10.1366/0003702824639231
Popis: Sensitive absorption measurements may be contaminated by the presence of particulates in the sample. Scattering by these particles can generate impulse noise (spikes) on the spectrum. This problem is especially serious in spectroscopy with a focused laser source, since the laser probes a small cross section of the sample. Often, this noise ruins an otherwise usable spectrum. If the data are stored digitally, however, it is often possible to remove noise spikes from a spectrum numerically. The commonly used smooth method of Savitzky and Golay does a poor job of removing these spikes. Since they are isolated and unidirectional errors, the smooth method simply transforms them to the impulse response of the filter function itself. Similarly, other low pass digital filters will also transform spikes into bands which may interfere with the analysis of real spectral features.
Databáze: OpenAIRE