Phase Unwrapping Algorithm Of Frequency Domain Optical Coherence Tomography

Optical Coherence Tomography(OCT)imaging technology is a non-contact detection technology that uses the principle of Michael interference to obtain the interference signal between the reference plane and the tomographic to be tested.It is a Non-invasive non-destructive tomography measurement.In the frequency domain optical coherence tomography measurement,the obtained interference signal is subjected to Fourier transform demodulation,and the phase information of the signal is obtained.The demodulation result is that the phase information of the measured internal sample corresponds to the tomographic plane of the material depth.However,with the Fourier transform of the interference signal,the phase information is wrapping between-? and ?.The wrapping phase needs to be unwrapping to obtain the depth information of the measured sample.The unwrapping result determines the measured accuracy of sample depth information.The noise in the measurement and the defects of the tested sample increase the difficulty of phase unwrapping.The stability of the phase unwrapping algorithm in the detection of the defect of the sample has become the key to the demodulation algorithm.In addition,in the actual measurement,the invalid area included in the field of view often needs to be manually removed,so the ability of the unwrapping algorithm to deal with the invalid area reflects the applicable range of the phase unwrapping algorithm.Two-dimensional phase unwrapping algorithms are mainly divided into two categories.One is to directly identify or fit the phase noise between wrapping phase and unwrapping phase,including the path following method and the optimization method.The other is to filter the wrapping phase,and then phase gradient and integration is used to unwrap the phase.This paper briefly introduces the optical path principle of optical coherence tomography system and the principle of demodulation algorithm.Phase unwrapping is an indispensable step in the processing of optical coherence tomography.Then,the phase unwrapping algorithm involved in the optical coherence tomography system is explored in detail,the principle of one-dimensional phase unwrapping is explored,and then the principle of two-dimensional unwrapping algorithm is explained.At the same time,the commonly used two-dimensional wrapping algorithm perform elaboration and comparative analysis.Path following of phase unwrapping is to limit the noise at a local range to prevent the cumulative effect of noise globally.It includes the branch cut method and the quality map method.The branch cut method is easy to cause pull lines and small islands.In optical coherence tomography of actual measurements,the choice of the appropriate quality map is very difficult;the optimization phase unwrapping method uses the gradient of the least squares to fit the phase gradient to unwrap,which can better retain the information of the material defect,but the phase unwrapping of the local area requires manual selection of the local area mask.In addition,in terms of phase filtering,the short-time Fourier transform filtering method will also distort the defect,but using the short-time Fourier transform can obtain the effective local area mask of the interference signal by selecting an appropriate threshold after the filtering by isolate ineffective or high noise areas.This paper proposes an automatic weighted least squares phase unwrapping algorithm.After filtering by short-time Fourier transform,the amplitude of the wrapping phase is used as a weighting matrix,and then the effective area is unwrapped using a weighted least squares algorithm.To preserve defect information of defect detection in the optical coherence tomography system avoid the influence of invalid area noise.By detecting the defect of the resin material,the optimized phase unwrapping method can obtain relatively smooth results.In addition,in the lens measurement experiment,the automatic weighted least squares unwrapping algorithm is used to reduce the effect of the defocused invalid area on the unwrapping result.