Splatting算法在三维CT迭代算法中的应用  

Splatting Algorithm Apply to Algebraic Arithmetic of Three Dimensional Computer Tomography

作　　者：莫会云[1] 潘晋孝[1] 

机构地区：[1]中北大学理学院电子测试技术国防重点实验室,太原030051

出　　处：《CT理论与应用研究（中英文）》2007年第3期6-12,共7页Computerized Tomography Theory and Applications

基　　金：国家自然基金资助(60532080)

摘　　要：采用代数重建技术重建三维物体,提出了一种基于splatting原理的ART权值计算方法,以增强重建图像的空间分辨率。该算法基于三维锥束扫描方式,采用双三次样条局部基函数作为插值核,并利用其足印函数对物体进行插值来计算像素的权值。与常量插值相比,这种方法对物体的灰度值有更精确的近似,在相同的实验条件下可重建出分辨率高的物体。给出了适用于三维锥束重建的splatting算法的具体实现,并对实验数据进行了实物重建及误差分析。This thesis uses ART(algebraic reconstruction technique) to reconstruct 3-D objects,Based on Splatting theory it presents a method of calculating ART weighting coefficients to enhance spatial resolution.This algorithm is based on the 3-D cone-beam scanning,using cubic B spline local basis function as interpolate kernel,and utilizing footprint function of filter kernel to interpolate the volume to calculate the weighting coefficients.Compared with the constant-size interpolation,the algebraic can gain more accurate object gray values.In the same experiment condition it can reconstruct high-resolution object.In the thesis,it gives the practical realization of splatting algebraic adapting to 3-D cone-beam reconstruction,and uses the experiment data to reconstruct the real object and analyze the error.

关 键 词：SPLATTING算法 代数重建技术 锥束CT 三次样条函数 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]