Parameters Calibration of CT System Based On Geometric Model and Inverse Radon Transform

Abstract

CT technology can make use of the penetrability of X-ray to obtain the internal information of the object without damaging the shape and structure of the sample. Therefore, it has a wide range of applications in the fields of industry, non-destructive testing and medical imaging. However, in the installation process of CT system, there are often some problems, such as inaccurate detector spacing, rotation angle and rotation center deviation, resulting in unclear image reconstruction. Thus it is necessary to calibrate the parameters of the installed system with template data, and then image the unknown samples in the process of CT system installation. This paper studies the parameter calibration of CT system. Through the pre-processed data, the analytic geometry model and statistical analysis method are established to calculate the detector interval, 180 rotation angles and rotation center of CT system. In order to verify the accuracy and stability of the parameters, Non-Local Means denoising algorithm (shortly known as NLM denoising algorithm) and Radon inverse transform imaging theorem are introduced. The theoretical image is rebuilt by the absorptivity of the template. The error between the theoretical image and the geometric information of the original template is analyzed, and the fitting is excellent. Two new inhomogeneous media are introduced, and the image is imaged by the parameters obtained from the parameter calibration model and the imaging theorem in this paper. The results are clear and the shadow points are excluded. At the same time, the specific absorptivity of the media is obtained. The calibration model of CT system parameters obtained in this paper, combined with imaging algorithm, can deal with different shapes and irregular conditions. It has great application value in the field of image reconstruction and medicine.