The Image Post Processing Technology Of Optical Coherence Tomography

Compared with other medical imaging techniques,optical coherence tomography(OCT)technology has the advantages of non-invasive,non-contact,high resolution,low cost and non-destructive,which is a rapid development of non-destructive of cross-sectional technique in recent years.It has been widely used in clinical diagnosis and high resolution detection of industry.However,as a kind of high resolution imaging technique,OCT is very strict to acquire the data of imaging data.Any section of system will inevitably affect the final imaging results,and absorption and scattering of light by high scattering media causes the phenomenon of resolution decline,lower imaging depth,low contrast and serious noise,that greatly limits the range of OCT application fields.Therefore,the post-processing of OCT image has a very important practical significance.Efficient and accurate algorithm will make the OCT system better applied to the actual diagnosis and detection.In this paper,from the perspective of digital image postprocessing,to improve the quality of OCT image as ultimate goal,we mainly focused on three parts of OCT image process: contrast enhancement,image denoising,image registration.Specific work content and achievements as follow:1)Aiming at low contrast of OCT images,combining with the characteristics of OCT image,we design and implement a contrast enhancement algorithm which based on the gamma correction.The algorithm has complementary advantages from global gamma correction algorithm and neighborhood gamma correction algorithm which has good universality and adjustability.Experimental results show that the algorithm not only can effectively enhance the image contrast and local details,speckle noise in the background can also achieve effective inhibition.Finally,we get better objective evaluation parameters than other methods.2)Aiming at the speckle noise in OCT images,by the analysis of the advantages and disadvantages of the traditional soft,hard and half threshold function,we propose an improved wavelet threshold denoising algorithm which combines the characteristics of the speckle noise and correct the threshold.The improved algorithm takes full account of the protection of image information during removing the speckle noise.The experimental results show that the algorithm can achieve better comprehensive evaluation results,and the quality of the processed OCT images is effectively improved.3)Aiming at the problem that denoising algorithm based on single frame will inevitably make the image clarity drop,back to the perspective of multiframe image processing,taking VS2012 as the development platform,we achieve OCT image registration before superposition average denoising processing with phase correlation algorithm and SURF registration function interfaces from OpenCV for continuous acquisition of OCT image data.The experimental results show that both registration algorithms can effectively improve the image SNR,make the image clearer,and the phase correlation method has faster speed and better denoising effect.