| Compared with traditional computed tomography(CT)systems,Cone-beam computed tomography(CBCT)imaging systems have advantages such as high spatial resolution,full utilization of X-ray radiation,and ease of operation.With the development of CBCT in recent years,researchers have gradually begun to study how to apply CBCT to soft tissue imaging fields,such as human thoracoabdominal imaging and small animal thoracoabdominal imaging.Compton scattering is one of the important factors affecting the quality of soft tissue imaging in CBCT.The Compton effect generates scattered photons,which interfere with projection data on detectors,increase background noise,and introduce scattered artifacts in the reconstruction process.This leads to decreased contrast and blurred details in the imaged object,which severely affects the contrast of low-density materials such as soft tissue.In addition,CBCT reconstruction algorithms are based on the assumption that X-ray energy spectra are monochromatic,but in reality,X-rays have a multi-energy spectrum,leading to errors in the reconstruction results and the introduction of hardening artifacts in the reconstructed image,manifested as shadows or highlights.Moreover,the inconsistent X-ray photon response of the flat-panel detector pixels introduces ring artifacts into the reconstructed image,severely contaminating the reconstructed image,deforming and even blurring and adhering tissue boundaries,further reducing the performance of soft tissue imaging.Therefore,in order to improve the contrast of soft tissue imaging in CBCT,special attention needs to be paid to issues such as scatter artifacts,beam hardening artifacts,and ring artifacts.Based on the attenuation plate(Beam Stop Array,BSA)for filtering scattered signals,this thesis proposes an algorithm to optimize the traditional interpolation algorithm..Compared with the traditional BSA method,this method does not require the radiation blocking point to completely block the amount of X-ray photons,and has lower requirements for the production of BSA.In this thesis,the scatter faction(SF)curve of the slit and BSA was measured,and the root mean square error(RMSE)of the curve was calculated.Using a head model as an example,the value of RMSE was 0.0083,which was in line with experimental expectations and could be considered accurate.In the beam hardening correction part,this study measured the Xray energy spectrum using an X123 cadmium telluride spectrometer produced by AmpTek,and analyzed the accuracy of the energy spectrum using an X-ray multifunctional dosimeter.Finally,based on the energy spectrum curve and detector response curve,a polynomial fitting method was used to correct the hardening artifacts.To verify the algorithm’s effectiveness,this thesis calculated the RMSE between the soft tissue reconstruction image of experimental animals’ thoracoabdominal regions and the fan beam CT(2mm slit).The RMSE value after applying this thesis’s correction method was 179.13 lower than that without correction,and the contrast noise ratio(CNR)increased by 31.90.In the ring artifact correction process,based on the characteristics of ring artifacts in the reconstructed image,this thesis proposed a oneway relative variation algorithm suitable for iteration in the polar coordinate system.Simulation results showed that after using this method proposed to remove ring artifacts,the structural similarity(SSIM)between the reconstructed image and the reference image was 0.97,and the RMSE value was 0.0017,which was close to the reference image.Experimental results have confirmed that after using the three methods mentioned above for scatter correction,hardening correction,and ring artifact correction,the soft tissue imaging ability of CBCT has been significantly improved. |
