提出一种基于蒂洪诺夫(Tikhonov)迭代的的电荷层析(electrostatic tomography, EST)成像算法。通过Tikhonov迭代法得到初始荷电粒子的图像分布,然后采用最大熵的方法设置门槛灰度对重建结果进行滤波提高图像的可分辨性。并针对Tikhonov迭代正则化系数难以选择的问题提出灵敏场奇异值分解的方法解决,选取灵敏场的最大奇异值作为正则化系数。仿真和实验结果表明:该算法具有收敛速度快,重建图像可分辨性高的优点。
An algorithm based on Tikhonov iterative for electrostatic tomography was presented. The first part of the scheme used the Tikhonov iterative algorithm to reconstruct the charged particles distribution, in the second part the entropic thresholding method was used to optimize the reconstruction image. For the problem of the regularization coefficient of Tikhonov iterative algorithm was difficult to decide, the singular value decomposition of the sensitivity was proposed and the maximum singular value is chosen as the regularization coefficient. Simulated data and experimental results indicate that the algorithm can provide high quality images with high convergence speed.