The Research Of The Polyenergetic X-ray CT Statistical Reconstruction Algorithm

Computed Tomography technique which has been widely used in many ways is one of the most advanced non-destructive detection techniques at present. It is especially used in industry inspection and medical diagnosis because the image reconstructed by CT are of good quality and high resolution.Traditional reconstruction algorithms assume that the X-ray is monoenergetic beams while in fact X-ray is polyenergetic either in industrial or in medical CT, so only polyenergetic projection data have been obtained in actual CT system. Once this kind of data is used to reconstruct the image directly, beam-hardening artifacts appear in the reconstructed image. If the object contains high absorption coefficient materials, the projection data is incomplete and abnormal as X-ray is absorbed, Once this kind of data is used to reconstruct the image directly, the metal artifacts appear in the reconstructed image. Additional, due to inconsistencies of projection data and reconstruction algorithm, beam-hardening artifacts and metal artifacts appeared.This paper based on statistical reconstruction algorithm, combining polyenergetic X-ray, constructing polyenergetic X-ray CT projection acquisition model and the object model, derived the polyenergetic X-ray CT statistical reconstruction algorithm. The projection data of the algorithm follow Poisson distribution, hypothesizing the object consists of a given number of non-overlapping known substances, the linear attenuation coefficient is the product of its unknown density and a known energy-dependent mass attenuation coefficient. Through the simulation and practical data experiment, the method in this paper can reduce artifacts effectively, reflect the original structure of object information, improved the image quality. But there are some shortcomings, the image is fuzzy around metal area and low contrast. In view of the above questions, and the algorithm was improved: using the image segmentation algorithm, obtained the segmentation image information, introduced image information for polyenergetic X-ray CT statistical reconstruction algorithm, and get the penalty function, strengthened the image edge and detail information, get the based on prior information polyenergetic X-ray CT statistical reconstruction algorithm. The experiment showed that, using this method, not only suppressing beam-hardening artifact and metal artifact, but also eliminating blurring around high-density substances and get high-contrast, improving image quality obviously.