A novel signal enhancement method for overlapped peaks with noise immunity. Issue 4 (20th April 2016)

• Record Type: Journal Article
• Authors: Li, Yuanlu; Pan, Chang; Xue, Yanmei; Meng, Xiao; Ding, Yaqing
• Is Part Of: Spectroscopy letters. Volume 49:Issue 4(2016)
• Page Start: 285
• Page End: 293
• Publication Date: 2016-04-20
• Subjects: Gaussian function -- noise immunity -- overlapped peaks -- signal resolution enhancement; Spectrum analysis -- Periodicals; 543.5
• DOI: 10.1080/00387010.2016.1144614
• Languages: English
• ISSNs: 0038-7010

Abstract

ABSTRACT: In order to detect the peak positions in overlapped spectrum, we usually need to sharpen overlapped peaks. In general, the second-order or fourth-order derivative is adopted to determine the peak position. However, as the derivative method is very sensitive to noise, the signal-to-noise ratio of derivative signal will be reduced and even affect the detection of characteristic peaks in the presence of noise. In this paper, a direct peak-sharpening method for the approximate of noisy overlapped peaks is proposed. First, the Gaussian second-order derivative function is sharpened, and then we conduct convolution operation between the sharpened signal and noisy signal and a smoothed second-order derivative sharpening signal is obtained. Meanwhile, the fourth-order derivative sharpening result can be easily obtained by apply sharpening again. The proposed method was verified by comparing several simulation peak groups. Moreover, the proposed method was used to enhance a mass spectrum signal and successfully located the position of each peak.