Abstract: The Landweber image reconstruction algorithm based on imaging sensitive field mean filtering method is proposed to solve the so-called “soft-field” characteristic problem of sensitive field of Electrical Capacitance Tomography（ECT） system. The algorithm principle is that the neighborhood pixel sensitivity is averaged by template convoluting method, which can reduce the sensitivity of region near electrodes and improve that of central region of pipe. The algorithm can eliminate partly the effect on the quality of image reconstruction due to uneven sensitivity and improve image reconstruction accuracy. Simulation results indicate that the algorithm is superior to conventional Landweber algorithm in image reconstruction accuracy and convergence speed.

Key words: Electrical Capacitance Tomography（ECT）, Landweber image reconstruction algorithm, sensitivity, mean filtering

摘要： 针对电容层析成像系统中敏感场的“软场”特性，提出了一种基于成像敏感场灵敏度均值滤波的Landweber图像重建算法。该算法是通过模板卷积的方式对敏感场灵敏度进行邻域平均，降低靠近极板区域的灵敏度，提高管道中心区域的灵敏度，部分消除了因敏感场不均匀对图像重建质量的影响，提高图像重建精度。仿真结果表明，该算法在图像重建精度和收敛速度上均优于传统的Landweber图像重建算法。

关键词: 电容层析成像, Landweber图像重建算法, 灵敏度, 均值滤波