3D-FLIT及其在工业CT体数据线特征提取中的应用

目的：特征检测是工业CT体数据处理中的重要问题，线特征是其重要特征之一，其检测效果将直接影响模式识别和分类。方法：有限线积分变换（finite line integral transform，简称FLIT）作为近年来新出现的一种多尺度几何分析方法，尤其适合于描述2D图像的线性奇异性。文中在2D-FLIT的理论基础上，将其推广到3D的情形，并结合图像融合和数学形态学的方法提出了两种基于FLIT的提取工业CT体数据中线特征的方法：一种是直接对工业CT体数据整体进行线特征的提取（3D-FLIT提取方法）；一种是先对工业CT体数据按某个方向进行切片划分，再对每张切片进行线特征的提取（2D-FLIT提取方法）。结果：实验表明，与小波的方法相比，上述两种方法能有效地提取体数据中的线特征。结论：该文的方法能对工业CT体数据中的线特征进行有效提取，为后续相关的数据处理（如逆向工程）打下坚实的基础。

3D-FLIT and its application to linear feature extraction of ICT cubic data

A linear feature extraction method for ICT cubic data was presented based on the Finite Line Integral Transform(FLIT),and the concept of the FLIT was introduced. Based on the concept,the 2D-FLIT was generalize into the 3D-FLIT to obtain the formulas of 3D-FLIT and inverse 3D-FLIT. By combining image fusion and morphological methods, two methods to extract the linear feature from cubic data based on FLIT (2D-FLIT or 3D-FLIT) were proposed.One of them is the idea of 3D-FLIT applying to the ICT cubic data dire...