Research On CT System Calibration Method Using Feature Texture Template

Computed Tomography can complete nondestructive detection for objects.It has advantages of high precision and easy to operation.Along with the rapid development of modern industry,medicine,security check,military and other domain,the using of CT has become more and more extensive.One of the most important evaluation index of CT is quality of the reconstructed images,which is major concern of the users.When the CT system couldn't satisfy the ideal geometric imaging relationship,the deviation of geometric parameters will make considerable influence on the reconstructed images.And this kind of influence will debase the quality of the reconstructed images.Hence,calibration of the CT system is essential.Different geometric parameters in CT system will make different influence on reconstructed images in CT systems.In this paper,the geometric imaging relationship of cone-beam CT and helical cone-beam CT is introduced.Secondly,The establishment of the coordinate system and the axis which shelving snit moves along in CT system is introduced.Thirdly,the deviation of geometric parameters is unified to the detector and the influence of different geometric parameters on reconstructed images is discussed.And the mathematical expression between different geometric parameters and coordinate error of back-projection is given is the part.For all geometric parameters of cone-beam CT and helical cone-beam CT,in this paper a new method of geometric parameters deviation detection is proposed.For the angle parameters deviation of detector,a kind of calibrate template is designed.This template contains of translational symmetric texture and rotational symmetric texture.Firstly,parts of angular deviations are solved through the rotational symmetry of calibration template.And then the translational symmetry texture is separated through frequency domain processing.The rest angular deviation is solved by the symmetry of translational symmetry texture.After the angular deviations of detector were solved,the detector is adjusted to make them to zero.Then operator makes the shelving unit move along the line linking ray source and detector and records the images collected by detector before and after moving.The displacement parameters are measured using the images combing the geometric imaging relationship of cone-beam CT.Finally,make the shelving unit move along the lift axis and record the images collected by detector before and after moving.Combing the geometric imaging relationship of helical cone-beam CT,the angular deviations of rotation axis and lifting axis can be measured.This method uses less projections during the measure process of the angle deviation of detector,and it can solve all angle deviation meanwhile.The operation of the method process is simple and the method can reach high precision.During the measure of displacement parameters and angle deviation between rotation axis and lifting axis,less manual intervention is needed.Meanwhile,the precision of feature point detection of the collected images is improved,which can improve the result of measure of displacement parameters and angle deviation between rotation axis and lifting axis.In order to verify the discussed result and the method in this paper,the simulation experiments and the actual experiments are designed.The simulation experiments results show that the discussion of influence of geometric parameters on reconstruction of images is correct.Then,the proposed calibration method is verified by simulation experiments and the actual experiments.