Contour segments extraction of 3-D surfaces from layer measurements

Abstract

By the layer-measuring methods of reverse engineering (RE) technology, such as industrial computed tomography (ICT), the geometrical data of a complex inner surface can be obtained. However, related researches, such as contour segmentation, 3-D model reconstruction, etc., need to be extensively carried out for industrial application. This paper intends to solve the problem of extracting contour segments corresponding to different types of surfaces/features from the layer-measuring data. In general, the extracting process can be divided into two stages: rough segmenting and refining. In the rough segmenting stage, an optimal contour matching method is put forward to find similar contour segments from another layer according to the seed contour. In the refining stage, different contour extracting methods are presented with respect to different types of surfaces. For polyhedral or rotational surfaces, an iterative way in which the feature contour can be extracted precisely is presented. For freeform surfaces, another approach is to detect the joining points between the target contour segment and others. The presented approaches can precisely extract open contour segments belonging to certain types of surfaces/features as demonstrated in the examples. Based on this paper, further research such as parameter identification of 3-D surfaces and CAD model creation can be conducted.