红外层析成像重构算法比较  

Comparison of Image Reconstruction Algorithm in Infrared Tomography

作　　者：杜亚举[1] 黄小濆[1] 邓迁 李伟[1] 李保生[1] 

机构地区：[1]合肥工业大学仪器科学与光电工程学院,安徽合肥230009

出　　处：《现代科学仪器》2014年第1期96-100,共5页Modern Scientific Instruments

基　　金：基金资助：国家自然基金面上项目资助（61174195）

摘　　要：光学层析成像具有分辨率高、无损伤性、便携性、连续性等特点，本文研究了光学层析成像图像重构两种常用的算法——直接傅里叶算法（DF）和滤波反投影算法（FBP）的理论方法。通过对在不同扫描投影角度数目下的图像进行模拟层析后，采用这两种算法进行了图像重构，比较了重构图像和原图像的均方差、重构时间和行像素值，结果证明DF算法的重构时间短、效率高，但重构图像的质量较低，而FBP算法重构结果的均方差低、质量好，但存在重构时间长、效率低的缺点。Optical Tomography has the features of high-resolution,no-injury, portability and continuity, this paper studies two commonly used image reconstruction algorithms in Optical Tomography, they are called as direct Fourier（DF） algorithm and filtered back projection （FBP） algorithm.Atier simulating the tomography process at different numbers of scanning projection angle,the 2-dimensional images were reconstructed by the two algorithms abovelThe mean square deviation,reconstruction time and line pixel value between reconstructed imades and original images were compared.The results showed that DF has short reconstruction time and high efficiency, but low quality, and FBP has low mean square deviation and good quality,but low efficiency.

关 键 词：光学层析成像 傅里叶 滤波反投影 图像重构 均方差 

分 类 号：TN216[电子电信—物理电子学]