Development Of μCT System For Small Animal Study

Medical imaging technology is the foundation of image processing technology and clinical application technology. The thesis is about the development of medical imaging equipment. Micro-CT imaging is an oncoming CT technology. With micron spatial resolution, it can get complex internal structure of the 3D information by completely nondestructive method with lower radiation and more exact reconstruction algorithm.The thesis uses Varian flat panel detector PaxScan1313 as the key component to build vivo micro-CT system for small animals. X-ray source and detector are fixed on a rack, which rotates with the electromotor when sampling. And still sample is put on a horizontal table fixed in the center of rotation. The operation of the whole system is controlled by the control center, a workstation. Control Centre controls the X-ray source, the data acquisition of detector. It also sends instructions to the motor driver, driving the X-ray source and detector rotating around the sample. FDK algorithm is utilized to reconstruct the image of specimens.In order to determine the relative coordinate between the point x-ray source and the detector plane, we invend a new method: the first step is to facture a tri-cone with transparent material, whose vertex is fixed with a metal ball. The second step is to rotate and to move the sample. So we can get several different projection images. According to each projection, we can list two equations. Through all the projections, we could have enough simultaneous equations including unknown numbers. And three unknowns determine the coordinate in space from the point x-ray source to the detector plane.In order to eliminate the cup-like artifacts caused by the effects of beam hardening in CT reconstruction, we find a new method of exponential fitting. The error of the exponential fitting is no more than 1/3 of the traditional two order polynomial fitting. We conclude that this method is more simple and precise than former classical fitting method.A new algorithm is presented for contour extraction in fan-beam CT reconstruction based on projection segmentation and backprojection. In most work on this problem, researchers segment two-dimensional slice images directly. Reconstructed slices are fuzzy, especially on the edges. Further segmentations operate on these fuzzy slices to get the contour results in more errors. The advantage of the new algorithm is its simplicity in segmenting projections. The results of simulation and experimental show its accuracy. This contour extraction method can save not only time and manpower, but also memory space.The amplification ratio of this micro-CT system is 2.2, so the special resolution is 58 micron. It can reconstruct the volume of a column with the diameter of 55mm and the height of 55mm. The repeat positioning accuracy is 0.005 degree. The step angle of rotation is 1.8 degree. We have used this micro-CT system to sample and reconstruct a trumpet shell and a Sprague Dawley rat. Three-dimensional rendering of results shows the system's utility.