Research On Ring Artifacts And Beam Hardening Artifacts Correction For Industrial Cone-beam CT Image

Due to differences between actual industrial cone beam Computed Tomography?CT?systems and the ideal system,there will be inevitably a variety of artifacts appeared in the reconstruction images,seriously degrading the image quality and further influencing the object recognition and analysis.Therefore,it is quite necessary to study effective industrial CT image artifact correction mehtods for improving the image quality and promoting the application of industrial cone beam CT.Correction methods for ring artifacts and beam hardening artifacts,the most common and serious artifacts in industrial cone beam CT images on information extraction and parameter calculation,are studied in this work.The main innovations are described as follows:1.A ring artifacts correction method based on linear regression for CT image is proposed.Traditional detector correction method depends on the pixels response only under static current cases to distinguish defective pixels and perform correction.The disadvantage is that defective pixels detection threshold is hard to be selected and there are always residual artifacts in CT images.Aiming for this problem,this thesis follows the physical characteristics that pixel response is linear to the tube current,and divides the pixels into defective pixels and inconformity pixels according to the pixel output characteristics under dynamic currents.Especially for the defective pixels which are difficulty to recognize under static current,a linear regression based method is employed.Through a fixed threshold the defective pixels are detected exactly.After interpolation,uniformity correction is implemented using the corresponding regression equation relevant to the average response of all pixels.Experimental results show that compared with the traditional method the proposed method is more thorough on ring artifacts suppression,and the signal to noise ratio of the image is improved.2.An effective-energy-spectrum-based beam hardening correction method for CT image is proposed.Among the existing hardware-based beam hardening correction method,filter based energy spectrum filtering method can block a large number of low energy photons that have great impact on beam hardening artifacts.However,due to lack of quantitative analysis on filter material and thickness,high and low energy photons in the filtered spectrum are out of tune.And it is difficult to achieve an effective beam hardening artifacts correction result.Aiming at this problem,according to both the attenuation coefficient of measured object,filter material,filter thickness,the initial energy spectrum and the filtered spectrum,a novel model for E_thre,the energy to be filtered,with the parameters of mean energy ratio and photon intensity integral is established.The energy spectrum could be effectively pre-hardened using the filter parameter obtained from E_thre.Experimental results show that the proposed method can not only reduce beam hardening artifacts,but also improve the signal to noise ratio of the image,and restore the contour and structure information of the object.3.A beam hardening artifacts correction method for local CT image is proposed.Because of a prominent advantage of saving hardware cost,local CT has widely application prospect.The beam hardening artifacts in local CT image could be weaken using a filter,but a software based method needs to be applied for further artifacts removement.However,affected by the projection data truncation,the existing polynomial based beam hardening artifacts correction methods cannot solve exactly the beam hardening correction parameters.Aiming at this problem,an exponential formed correction model is established for beam hardening correction parameters calculated,and let the minimum gray entropy of the image as the goal in solving the parameters.To reduce the effects of the bright ring-like truncation artifacts caused by the truncated projections on entropy values,a Radon inversion transform-based interior reconstruction method is adopted and the parameters could be calculated accurately.Experimental results show that the improved polynomial based beam hardening artifacts correction method can overcome the interference of projection truncation,effectively correct beam hardening artifacts in local CT reconstruction images and eliminate bright ring-like artifacts caused by truncated projection data.