The Analysis And Elimination Of Optical Beat Noise In Fiber Low-coherence Interometry

Low-coherence interferometry is popularly used as a high-sensitivity measuring tool in science and industry. Low-coherence interferometry can measure optical phase with reducing coherence noise and moreover can carry out the absolute optical path difference, including optical coherence tomography. A low-coherence interferometer is illuminated by a broadband light, which is generated by spontaneous emission from various incoherent light sources. With the increasing use of optical coherence tomography, the study of the signal-to-noise ratio in low-coherence interferometry has been of increasing concern. The particular noise in low-coherence interferometry is the optical beat noise due to the mixing of the multiple frequency components involved in the spontaneous emission of an illuminating low-coherence light. At a certain range the optical beat noise will limit the optical path difference and optical phase detection sensitivity.This thesis makes a detailed theoretical study of the sensing performance of low-coherence interferometry, and proposes the basic formulas of the thermal noise, the shot noise and the optical beat noise. Because of the small phase change and absolute optical path change, this thesis presents the basic formulas of the signal to noise ratio and resolution. The significant effect of the optical beat noise on the performance of low-coherence interferometer are numerically evaluated and discussed.LabVIEW virtual instrument is introduced. At LabVIEW platform the interference signal is processed, including data acquisition, display, low-pass filtering, median filtering, signal to noise ratio estimation and the beat frequency noise extraction. Powerful frequency-domain function modules in LabVIEW formed a very good complementary with MATLAB wavelet toolbox. This thesis mainly analyzing interference signal, based on traditional filtering method, fast Fourier filtering method and wavelet threshold denoising method. Through large experiment results comparing analysis, proves their each excellences and flaws. Multi-resolution analysis character of wavelet transform can well reflect signal local characteristic. The selection of wavelet threshold and wavelet base are the emphasis of this thesis.