在计算机层析技术中投影数据的多目标优化  

Multicriterion Optimization for Projection Data in Computerized Tomography

作　　者：李虹 李宁[2] 王旭东[2] 李时光[2] 郑尔信[2] 

机构地区：[1]重庆电视台,重庆400039 [2]重庆大学,重庆400044

出　　处：《重庆大学学报（自然科学版）》2001年第1期67-69,83,共4页Journal of Chongqing University

摘　　要：在计算机层析技术 (CT)中 ,优化投影数据是改善重建图象质量的有效途径。为减弱噪声等因素对投影数据所造成的模糊 ,也为了寻求最佳唯一图象 ,为此以模糊数学和决策理论为基础 ,通过建立投影模糊指数函数及平方误差模糊指数函数 ,提出了一种称为多目标优化的新的模型 ,对投影数据进行优化。实验验证在微机上完成 ,首先在给定图象上作仿真采集并将所得投影数据加高斯噪声 ,然后用含噪声的及经多目标优化的两类投影数据分别完成图象重建 ,并把所得结果作对比分析。实验结果表明 ,所提出的多目标优化模型有较强的抗噪声能力 ,理论和实验有较好的一致性。It is an effective way for improving quality of reconstructed images that projection data are optimized in the field of Computerized Tomography (CT). For lowering the fuzziness on projection data caused by lots of factors such as noise and getting the unigue optimal reconstructed image from noisy projection data further more, in this paper, the authors develop a new model named Multicriterion Optimization Model (MOP) for optimizing projection data. The model is structured based on the theories of fuzzy mathematics and decision making, via the method of deducing projection fuzzy exponent function and square error fuzzy exponent function. The experiments for validating the model have been carried out on personal computer (PC). At first, we make simulation collecting in the images presented and add Gauss Noise into the projection data obtained by simulation collecting, Then, complete the image reconstruction from noisy projection data by the solution without optimization and another solution with multicriterion optimization. Finally, compare and analysis the different results about images reconstructed by two different solutions. The experiment results indicate that the MOP in this paper has better consistency with the theory and practice as well as obvious advantage of antinoise ability.

关 键 词：计算机层析 噪声 投影数据 模糊指数函数 多目标优化模型 

分 类 号：TP391.5[自动化与计算机技术—计算机应用技术]