Study On The OCT Image Enhancement And Calibration Model Of Non-invasive Blood Glucose Monitoring

Optical coherence tomography(OCT)is a new image technology with high resolution,which has been widely used in biomedical and material testing.OCT can be used for noninvasive blood glucose monitoring by measuring the optical parameters of the tissue fluid affected by the change of human blood glucose concentration(BGC).This paper mainly studies the correlation analysis model of OCT non-invasive blood glucose detection.In addition,an image enhancement algorithm is proposed to improve the image visualization and signal to noise ratio(SNR).In this paper,according to the relationship between the skin non-uniformity and the relationship between the scattering coefficient and BGC,we propose a two-dimensional correlation calibration model based on the one-dimensional calibration algorithm.Through the model,the distribution of correlation coefficient(R)between the scattering coefficient and the BGC of the two-dimensional skin cross section can be get.Thus we establish R-filter to select skin area with a relatively high R of and eliminate the low R.Then the optimal calibration region can be filtered by calculating the overall relevancy(OR)between the scattering coefficient and BGC.The model was confirmed by non-invasive blood glucose monitoring experiment.The results show that the two dimensional correlation calibration model is superior to the one dimensional calibration algorithm.Finally,based on the two-dimensional correlation analysis model,the spatial three-dimensional correlation analysis method is proposed with higher accuracy in theory.This paper studies the influence of signal attenuation and high scattering on OCT images,and proposes an image enhancement algorithm based on differential processing.We set an experiment to validate the algorithm,and the results show that this algorithm can improve the image visualization,and eliminate the influence of signal attenuation of the image,which has potential application value for OCT non-invasive blood glucose detection.