基于改进果蝇优化算法的ECT图像重建方法

An Image Reconstruction Method of ECT Based on Improved Fruit Fly Optimization Algorithm

电容层析成像(ECT)的反问题具有严重的不适定性,因此经典算法的重建结果往往精度不足,并伴有很多伪影。为了提升重建图像质量,考虑将果蝇优化算法(Fruit Fly Optimization Algorithm, FOA)应用在ECT图像重建上。FOA作为群智能搜索算法,具有原理简单、参数设置少、容易实现等优点,但是容易陷入局部最优。针对该问题,提出了一种改进果蝇优化算法,通过与Tikhonov正则化算法相结合求解ECT图像重建问题。为了验证新算法的有效性,选择六种流型作为重建目标,并将LBP算法、Tikhonov正则化算法以及Landweber算法的重建结果作为对照组。仿真实验结果表明,本文算法可以有效提高图像重建精度,重建图像在相对误差和相关系数方面均有所提升,验证了该算法的有效性。

Electrical capacitance tomography(ECT)inverse problem has serious ill-posed feature, therefore the reconstruction results of classic algorithms are often inaccurate and have much artifacts. In order to improve the quality of reconstructed images, the Fruit Fly Optimization Algorithm(FOA)is considered to apply in image reconstruction of ECT. As a swarm intelligent search algorithm, FOA has the advantages of simple principle, few parameters, and easy implementation, but it is easy to fall into local optimum. To overcome this disadvantage, an improved FOA,which combined with Tikhonov regularization algorithm, was proposed to solve the ECT image reconstruction problem. In order to verify the effectiveness of this algorithm, six flow patterns are selected as reconstruction targets, and the reconstruction results of LBP algorithm, Tikhonov regularization algorithm and Landweber algorithm are employed as compared objects. The results of simulation experiments imply that the proposed algorithm can promote the imaging accuracy and the relative errors and correlation coefficients of the reconstructed images are also improved, which illustrate the effectiveness of this algorithm.