两约束条件正则化的ECT图像重建算法

为提高图像重建的性能,提出了一种以灰度能量最小和图像二阶光滑性为约束条件,合并单位矩阵和位置相关二阶微分算子矩阵,构建正则化矩阵的电容层析成像重建算法。新的正则化算法与目前常用的标准Tikhonov正则化算法不同在于目标函数中的正则化项约束水平随图像单元位置变化,达到在整个成像区获得光滑一致的效果。仿真结果表明,新的算法与标准算法相比较,其重建图像性能得到了改善和提高。

Two Combined Restriction Regularization Algorithm for Electrical Capacitance Tomography Visualization

Based on the restrictions of minimum of gray energy and second order smoothness,the paper presented.an novel regularization method for electrical capacitance tomography by which a regularization matrix was constructed using the combination of an identity matrix and a second order differential operator.The difference between the new algorithm and the common regularization ones is that the regularization levels related to the site of the pixels vary spatially to achieve the consistently smoothing effects in different sites of the region imaged.Simulation experimentations have been performed to verify that the new algorithm is capable of reconstructing superior image quality over the standard Tikhonov algorithm.