Parallel Spectral Domain OCT System Design And Image Processing Algorithm Research

Optical Coherence Tomography(OCT)is a high-resolution,non-destructive testing technology based on the principle of optical low coherence,which is often used in the fields of biomedical imaging and industrial non-destructive testing.In order to solve the problem that the imaging speed and phase stability of traditional spectral domain OCT(SD-OCT)systems are limited by mechanical scanning,this paper proposes a parallel SD-OCT system based on line light source.Parallel SD-OCT system does not require any mechanical scanning.It uses area array CMOS camera and line source to complete the two-dimensional cross-section direct imaging,and on this basis,completes the optimization of the system’s key performance parameters and the removal of image speckle noise.1.Completed the optical simulation,structural optimization,construction and debugging of the free space parallel SD-OCT system with a center wavelength of 850 nm,and introduced the data processing of the system,and performed a simulation analysis for each processing process.It was successful complete the performance measurement experiment of the system and the two-dimensional cross-sectional imaging experiment of glass and coins.After final debugging,the parallel SD-OCT system has an axial resolution of 11.72μm,a maximum sensitivity of 20.25 d B in air,an imaging depth of1 mm,and a lateral detection range of 8.55 mm.2.In order to improve the key indicators of the parallel SD-OCT system(Imaging depth,Axial resolution,Sensitivity),optimization algorithms are proposed from three aspects.Aiming at the wavelength calibration in the resampling process,this paper proposes two different calibration methods based on mercury-argon lamp and based on interference spectroscopy,and completed comparative experiments.The results show that both methods can improve the performance of the system.Aiming at the problem that the sensitivity of parallel SD-OCT decreases with depth,this paper proposes an image reconstruction algorithm based on non-uniform discrete Fourier transform(NDFT),which not only improves the axial resolution,but also increases the sensitivity of the system to a maximum of 4d B.Aiming at the problem that the quality of the reconstructed image of the parallel SD-OCT system is affected by the point spread function of the light source,an image deblurring algorithm based on Richardson-Lucy deconvolution is proposed,which not only improves the definition and contrast of the reconstructed image,but also maximizes the sensitivity of the system.The increase has also reached 7d B,and the bottom widening of the axial point spread function has also been reduced.3.This paper proposes two speckle noise removal algorithms: anisotropic processing(AD)denoising、dual-tree complex wavelet transform(DTCWT)threshold denoising.Propose a dual-tree complex wavelet algorithm based on an improved threshold function,and complete related denoising experiments.Compared with wavelet transform,dual-tree complex wavelet transform has lower preservation redundancy,unbiased frequency,preserves more image details and phase information,and has better denoising effect.The AD algorithm can well maintain the edge details of the image while denoising.Although compared to other algorithms,both the AD algorithm and the DTCWT algorithm can maintain high-frequency details while denoising,the AD algorithm is better than the DTCWT algorithm in maintaining edges,and the DTCWT algorithm is better than the AD algorithm in removing noise.