baselineCorrectionCWT: background fitting via penalized least squares algorithm

In zmzhang/baselineWavelet: baseline correction by wavelet and Whittaker Smooth algorithms

Description

intelligent background-correction algorithm is developed, which simulates manual background-correction procedure of an expert intelligently. It basically consists of works of three aspects: 1) the accurate peak position detection of Raman spectrum by continuous wavelet transform (CWT) with the Mexican Hat wavelet as the mother wavelet; 2) peak widths estimation by enhanced signal-to-noise ratio (SNR) derivative calculation based on CWT but with the Haar wavelet as the mother wavelet; and 3) background fitting using penalized least squares with binary masks.

Usage

baselineCorrectionCWT(x,peakWidth,threshold=0.5,lambda=100)

Arguments

x	Raman spectrum
peakWidth	returned by widthEstimationCWT
threshold	the user define peak shape threshold
lambda	lambda is an adjustable parameter, it can be adjusted by user. The larger lambda is, the smoother fitted background will be
differences	an integer indicating the order of the difference of penalties of Whittaker Smoother method, see WhittakerSmooth

Value

the final background

Author(s)

Yizeng Liang ,Zhang Zhimin

See Also

WhittakerSmooth

Examples

data(raman)
x=m[7,]
scales <-seq(1, 70, 1)
wCoefs <- cwt(x, scales=scales, wavelet='mexh')
image(1:nrow(wCoefs), scales, wCoefs, col=terrain.colors(256), axes=FALSE, xlab='index', ylab='CWT coefficient scale', main='CWT coefficients')
box()
localMax <- getLocalMaximumCWT(wCoefs)
ridgeList <- getRidge(localMax, gapTh=3, skip=2)
plotRidgeList(ridgeList)
majorPeakInfo = identifyMajorPeaks(x, ridgeList, wCoefs, SNR.Th=1,ridgeLength=5)
peakWidth=widthEstimationCWT(x,majorPeakInfo)                                                
backgr = baselineCorrectionCWT(x,peakWidth,lambda=1000,differences=1)
corrected=x-backgr 
plot(xa,x,type='l',ylim=c(min(c(x,corrected)),max(c(x,corrected))),main="The background-correction result of Raman Spectra",xlab=expression("Wavenumber / cm"^-1),ylab="Raman Intensity/Arbitr. Units")
points(xa[majorPeakInfo$peakIndex],x[majorPeakInfo$peakIndex])
lines(xa,backgr,lty=5)
lines(xa,corrected)

zmzhang/baselineWavelet documentation built on Dec. 26, 2019, 8:49 a.m.