Geometric Calibration Method For High Resolution Cone-beam CT System

With the development of sub-micron focused X-ray sources and high-resolution detector technology and the availability of mature products,high-resolution cone-beam CT imaging technology has achieved rapid development.Micro-nano CT refers to cone-beam CT with spatial resolution of several tens to several hundreds of nanometers.It adopts a panel detector ray-beam projection scanning and direct three-dimensional reconstruction imaging.It has the characteristics of high resolution and fast imaging speed,with a wide range of applications.The spatial resolution of micro-nano CT system is affected not only by the X-ray source focus and intrinsic resolution,but also has a close relationship with the geometric parameters of the system.During the initial installation and long-term use of the CT system,deviations in the geometric parameters of the system and focal spot drift of the X-ray source will lead to distortion or blurring of the reconstructed image and reduce the spatial resolution of the image.Therefore,it's important to study how to calibrate the geometric parameters deviation and the focal spot drift of the ray source.In this paper,the high-resolution cone-beam CT system is taken as the research object,and the system geometric parameter deviation correction method and the focal spot drift correction method are studied.The research contents and innovations of this paper are listed as follows:Firstly,the influence of geometric parameters deviation on reconstructed images is analyzed through theoretical derivation and simulation experiments,and an improved method based on the square-line phantom is proposed.Acquire the square-line phantom projection at two different radial positions of the detector,the source-detector-distance,the source-rotation-distance and the vertical position of the detector could be determined by processing these projected images.Calculate the horizontal position of the detector by analyzing the square-line phantom projection at the angles with 180 degrees apart.Resolve the remaining geometric parameters by using the analysis results of the model and it's projected images.Finally reconstruct images using modified FDK algorithm.The proposed method can solve geometric parameters accurately and correct the reconstructed image artifacts effectively.The results verify phantom noise immunity,installation simplicity and practicality.Secondly,discuss the influence of focal spot drift on reconstructed image quality.And investigate the factors that affect the focal spot drift through actual experiments.Experimental results show that the lab temperature has little effect on the focal spot drift,while the power of the X-ray source can seriously affect the drift of the focus,and the focal spot drift increases with the increase of power.Thirdly,this paper presents a reference scan compensation method for calibrating the drift of the focus.A reference scan is performed immediately after the actual scan,and these projected images are compared by the extended phase correlation method to solve the focal spot drift.Then compensate the object projections by using the obtained drift,and reconstruct the projections.The experimental results demonstrate that this method can accurately solve the focal spot drift and effectively improve the quality of reconstructed images.