为了提高基于工业CT图像重构的三角网格质量，提出了顶点调整和特征因子相结合的网格规范化算法。引入模型特征因子和局部网格质量提高程度作为网格调整的控制条件，保留原始模型的局部细节特征; 采用法矢量阈值按规律递增的方式自适应控制网格调整，实现不同曲率特征的自动识别。实验结果表明，与现有方法相比，该方法能更好地规范模型的三角网格，同时保留了原始模型的细节特征。

In order to improve the triangular mesh quality of the mesh models reconstructed by industrial CT slice images, the mesh regularization algorithm combining the vertex adjustment and feature factors was presented. Introducing the improvement degree of the local mesh quality and the model feature factors as the control condition of the mesh adjustment, the local detail features of the original model were retained. The way of self-adaptive control mesh adjustment was adopted according to the normal vector threshold increasing regularity, different curvature features were identified automatically. Compared with the existing methods, experimental results show that the presented method can better regulate the triangular mesh while preserving the details features of the original model.