Jitter Correction and SNR Improvement of A-Scan Signal Using Wavelet Denoising and Signal Averaging for a Portable Time-Domain Optical Coherence Tomography System

Abstract

<p class='IJASEITAbtract'>Optical coherence tomography (OCT) is used to probe surface cross-sections of various materials in manufacturing, anatomy, and agriculture, among others. In any OCT system, signals have embedded noise in various forms, such as electrical noise and Jitter, which affect the depth profile (A-scan) and image quality. A signal processing method for correcting the Jitter and improving the signal-to-noise ratios (SNR) and image quality was developed in this study for a portable time-domain (TD) OCT system that utilizes a pc-based oscilloscope for data acquisition. A stack of five glass coverslips was used as a sample. Each signal from the oscilloscope consists of an A-scan coupled with a trigger signal. Jitter correction was done by first denoising the trigger signal using a combination of moving average and wavelet denoising. A reference trigger was selected, and all other trigger signals were adjusted, including the A-scans. Once jitter was corrected, the denoising method on the OCT A-scans was employed. The signal averaging method and various wavelet denoising methods were applied to the A-scans of the sample to identify which will give the highest SNR and improved image quality. Combining averaging of fifty signals and Daubechies 7 (Db7) with hard thresholding reduced the acquisition time and storage space by 50%, improved the SNR by 18 dB, improved the depth profile and image quality of a stack of five glass coverslips. This signal processing method will allow us to characterize and properly visualize cross-sectional images of other samples in the future using our TD-OCT system.