Algorithm Research Of Image Reconstruction Based On Circular X-ray Three-Dimensional CT

Cone-Beam CT is one of the most advanced subjects on the field of CT. Compared with 2D fan-beam CT it can achieve higher spatial resolution, higher scan speed and higher efficiency of X-ray. These advantages are very important in medical and non-destruction.First of all, we introduce the cone beam exact reconstruction algorithms, summarize the work which were done before. At the same time, we explain the conditions of them. Then we describe and analyze three approximate reconstruction algorithms carefully, the FDK method, crossing convolution method, conical convolution method. After that we mention the advantages and disadvantages of every one.The methods of calculating the cone beam projection of two typical common 3D geometries are deduced. Furthermore, we also introduce the simulation of digital projections. The cause of noise in the synthesized projections is analyzed. According to improvement and computer simulation, we compare the effect of them.Based on the characteristic of industrial flat-panel detector, in order to improve the quality of the image reconstructed with FDK method and restrain its noise, some research about improved FDK method and projection's disposal have been mentioned in this paper. From the problem how the noise of projection happen, we give improvement about the FDK method. Using the length which traverse the voxel for the coefficient of back projection value, after simulation we obtain better effect. It can eliminate the noise and increase the image quality on the x-y plane and across z direction, but the speed of computation will slow down.