基于正则代数重建算法的二维温度场重建

Two-Dimensional Temperature Field Reconstruction Based on Regular Algebraic Reconstruction Technique

在使用代数重建算法(Algebraic Reconstruction Technique, ART)对二维非均匀温度场进行重建时,离散误差和投影噪声会随着迭代修正被引入,为了平衡离散误差,减小算法对噪声的敏感度,在ART中引入了正则化项,并使用留一交叉验证法对单位正则化参数进行了选取,根据投影穿过待测区域路径的长度和单位正则化参数动态调整每条投影的正则化权重,实现了对每条投影离散误差和噪声水平的衡量。在不同的投影分布情况下,使用该算法对高斯单峰对称和高斯单峰偏置温度场进行了仿真重建,重建结果表明相比于传统迭代算法,该算法可有效提高温度场的重建精度,并且具有较好的稳定性。

When using the algebra reconstruction technique(ART)to reconstruct a two-dimensional non-uniform temperature field,the discrete error and projection noise are introduced with iterative correction.In order to balance the discrete error and reduce the sensitivity of the algorithm to noise,a regularization term is introduced into the ART algorithm,and the unit regularization parameter is selected by using the left-over cross validation method.The regularization weights of each projection are dynamically adjusted according to the length of the projection path through the area to be measured and the unit regularization parameters,the discrete error and noise level of each projection are measured separately.Under different projection distributions,the algorithm is used to simulate and reconstruct the Gaussian single peak symmetric and Gaussian single peak biased temperature fields.The reconstruction results show that compared with traditional iterative algorithms,the algorithm can effectively improve the reconstruction accuracy of the temperature field and has good stability.