计算机断层成像（CT）技术广泛应用于工业、医疗、国防、航空航天等领域。随着科学技术的不断发展，对使用CT来获得更加精确的材料内部结构信息提出了更高的要求，希望能在重建图像中更加完整地表达组织结构信息。为了获取更好的图像质量，提出一种基于指数形式的滤波函数（EBFF），在MATLAB平台下采用EBFF对图像进行重建，并通过相似系数、归一化平均值以及归一化均方根距离等评价手段与几种传统的滤波函数重建图像结果进行对比，同时在锥束CT实验平台下采用EBFF进行图像重建。模拟与试验均表明EBFF对重建图像质量有一定改善，其较传统滤波函数能给出更好的重建结果，有助于图像重建结果的优化。

Computed tomography (CT) has been widely applied in industrial, medical, defense, aerospace and other fields. With the continuous development of science and technology, higher requirements on CT for getting more accurate information on the internal structure of the material are put forward and one hopes to display more detailed and complete organizational structure information in reconstruction image. In this article, an exponential-based filter function(EBFF) is proposed to obtain a better image quality by using exponential-based filter function to reconstruct the image in the MATLAB platform, and the parameters of ε, r and d are exploited to compare image reconstruction results with several traditional filter functions. Furthermore, EBFF was used to reconstruct the image under the experimental platform of cone beam CT. Both simulation and experiment results show that EBFF has some improvement on the reconstructed image quality, and compared to the traditional filter function EBFF can give a better reconstruction result and is helpful to the optimization of the results of image reconstruction.