基于PCA降维和迭代正则化的温度场重建  

Temperature Field Reconstruction Based on PCA Dimension Reduction and Iterative Regularization

作　　者：白云 颜华[1] 魏元焜 BAI Yun;YAN Hua;WEI Yuankun(School of Information Science and Engineering,Shenyang University of Technology,Shenyang 110870,China)

机构地区：[1]沈阳工业大学信息科学与工程学院,沈阳110870

出　　处：《自动化与仪表》2023年第9期21-26,共6页Automation & Instrumentation

基　　金：国家自然科学基金项目(61372154);辽宁省博士启动基金项目(201601157)。

摘　　要：针对声学CT对温度场重建的不适定性,该文提出了一种基于PCA降维和迭代正则化的重建算法。通过径向基函数逼近声慢度分布建立声学CT正问题模型;用PCA降维改善逆问题的病态性;用迭代正则化法求解逆问题;利用声慢度与温度的关系得到温度分布。仿真和实际温度场重建实验表明,与常用的最小二乘法和基于奇异值分解的直接正则化法相比,所提算法的重建图像更接近真实分布,重建误差最高可降低86.62%和29.1%。因此基于PCA降维和迭代正则化的重建算法能够提供更高质量的重建温度场。Aiming at the ill-posed nature of acoustic CT for temperature field reconstruction,a reconstruction algorithm based on PCA dimension reduction and iterative regularization is proposed.The forward problem model of acoustic CT is established by approximating the sound slowness distribution with radial basis function.The ill-condition of the inverse problem is improved by the PCA dimension reduction.The inverse problem is solved by an iterative regularization method.The temperature distribution is obtained by using the relationship between sound slowness and temperature.Simulation and actual temperature field reconstruction experiments show that compared with the least square method and the direct regularization method based on singular value decomposition,the reconstructed image of the proposed algorithm is closer to the real distribution,the reconstruction error can be reduced by 86.62%and 29.1%at most.Therefore,the reconstruction algorithm based on PCA dimension reduction and iterative regularization can provide a higher quality reconstruction temperature field.

关 键 词：声学CT 温度场重建 重建算法 PCA降维 迭代正则化 病态逆问题 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]