A NEW PRINCIPLE ON 3D-RECONSTRUCTION FROM A FEW PROJECTIONS OF SLANT SCANNING.

本文研究了一种基于少量倾斜扫描投影的快速三维重建方法。该方法的重点是利用奇异值分解和采样模型对象，从少量倾斜扫描投影中快速重建三维图像。首先，我们研究了一种新的奇异值分解算法。基于该算法的计算机仿真实验结果良好。得到了重构参数间的良好条件。其次，我们尝试设计一种微小的倾斜扫描机构和x射线设备的软件接口，并成功地获取了一些投影图像。第三，对重建算法进行了实验，成功实现了由少量倾斜扫描投影进行三维成像。重建的三维图像效果良好。几次实验清楚地表明，这种重建方法有可能进入实际应用。

The study is concerned with developing a fast 3D reconatruction method from a few projections of slant scanning. The point of this method is rapidly to reconstruct 3D images from a few projections of slant scanning, using the Singular Value Decompositions and the sampling model object. The first, we investigated a new algorithm by using the SVD.The results of computer simulation experiments based on the algorithm were good. We obtained good conditions between reconstructing parameters. The second, we tried to design a minute mechanism of slant scanning and software interface with a X-ray equipment, and succeeded in acquisition of a few projecting images. The third, we experimented with the reconstruction algorithm, and succeeded in 3D imaging from a few projection of slant scanning. The 3D images reconstructed were good. Several experiments made it obvious that there is a possibility that the reconstruction method will come into practical use.