Gaussian窗函数在电容成像图像重建中的应用

为降低敏感场矩阵的小奇异值给电容成像(electri-cal capacitance tomography,ECT)问题带来的不稳定性,以提高图像重建质量,该文从正则化方法的谱滤波形式出发,提出了一种基于Gaussian窗函数的ECT图像重建正则化方法。该方法利用Gaussian窗函数的滤波特性,通过选择合适的Gaussian窗函数幂次,来达到降低敏感场矩阵的小奇异值对重建结果造成不稳定的效果,从而获得较好的图像重建结果。对5种典型介电常数分布进行了数值仿真,结果表明:该算法对于两相流不同介质分布均具有良好的适应性,可获得较好的成像结果,部分分布的重建图像轮廓较Tikhonov正则化方法更为清晰。

Gaussian window function for image reconstruction for electrical capacitance tomography

A reconstruction algorithm based on the Gaussian window function was developed to improve the quality of reconstructed images for electrical capacitance tomography.The system uses spectral filtering with the proper power of the spectral filter based on the Gaussian window function,measurement error instabilities and the small singular values of the sensitivity matrix.Simulations using five typical permittivity distributions provide reconstruction results that are better with some clearer contours than Tikhonov regularization.The numerical results verify the feasibility of the image reconstruction algorithm using the Gaussian window function for regularization.