Research On Image Reconstruction Algorithm Based On Optical Tomography Scanning System

Tomography technology is widely used in many fields such as nanometer tomography,neutron tomography,aerospace precision instrument detection and geological material detection because of its non-contact,long-distance and high-precision imaging characteristics.Due to the different application scope and structure design of tomography system,the projection data may be incomplete or deviated in actual acquisition,so that the reconstructed image will produce artifacts and unclear problems.Aiming at the problem of artifacts in reconstructed image in optical tomography system,a series of researches on image reconstruction algorithms were done.In order to improve the resolution of reconstructed image,an improved iterative algorithm was proposed by analyzing the filtering back projection algorithm and iterative algorithm.The main research work of this paper is as follows:The first step was to study the theoretical mechanism of the optical tomography system.There is a problem of systematic center offset and incomplete data collection in data acquisition.Then,center correction and data supplementation were performed on the problem to implement data preprocessing and provide accurate projection data for reconstruction.The influence of several parameters on the iteration effect was analyzed in the traditional iteration algorithm and reconstruction process.Then by analyzing we selected the appropriate weighting model,iteration times and relaxation factors.An improved ART algorithm and an improved SART algorithm were proposed through combining with high quality initial image and convex set constraints.A simple optical tomography system was used as an experimental platform according to the theoretical knowledge of imaging system.The projection data of the measured objects with different characteristics were collected and preprocessed.Then,through selecting different iteration times or relaxation factors for iterative reconstruction,the appropriate parameters were obtained by comparing with their results.The improved ART algorithm,SART algorithm and filter back projection algorithm were used to reconstruct the image separately.The improved iterative algorithm could reconstruct the image with high resolution by using image evaluation parameters such as sharpness and average gradient.