A NEW APPROACH FOR SPECTRA BASELINE CORRECTION USING SPARSE REPRESENTATION

摘要

A new baseline correction algorithm for spectral signalrnbased on sparse representation is proposed. Firstly, utilizingrnthe training sample to obtain the dictionaries of bothrnbaseline and spectrum; Secondly, establishing sparse representationrnmodel of spectral signal; thirdly, employing OMPrnalgorithm to calculate the representation coefficients ofrnspectral signal and finally, obtaining the spectral baselinernfrom representation coefficients which are corresponded tornthe baseline dictionary. Then, the spectra baseline correctionrnis completed by removing the baseline from originalrnobserved spectrum. Contrast experiment and quantitativernanalysis of corrected spectral signals are conducted and resultsrnshow the highly efficiency and accuracy of the proposedrnalgorithm.