Point cloud segmentation for complex microsurfaces based on feature line fitting
- PDF / 4,031,856 Bytes
- 26 Pages / 439.37 x 666.142 pts Page_size
- 96 Downloads / 177 Views
Point cloud segmentation for complex microsurfaces based on feature line fitting Xiaogang Ji 1,2
1
& Xixi Zhang & Haitao Hu
1
Received: 4 April 2020 / Revised: 2 September 2020 / Accepted: 16 September 2020 # Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract
Surfaces based on feature line constraints have higher accuracy than free-form surfaces and can capture other geometric relations of the model. The parts of complex microsurfaces are formed by arrays and crossings of several small surfaces. Many problems can be encountered in identifying feature points and fitting feature lines, which are difficult to solve by reverse engineering. In this study, feature point extraction, feature line fitting, and three-dimensional segmentation were investigated. First, the connection between two surfaces and the corresponding differential geometric quantities were explored. Then, a feature point extraction method for complex models was proposed. Second, the problems of separation, simplification, and combination of feature points for different models were analyzed, and the feature lines used to segment the point cloud were constructed. Finally, a region growth method based on feature line constraints was proposed to segment the point cloud data. Experimental results show that this method can solve the problem of excessive and insufficient segmentation for complex microsurface point cloud data and thus represents a foundation for high-quality model reconstruction. Keywords Reverse engineering . Feature point extraction . Feature line fitting . Point cloud data segmentation . Model reconstruction
1 Introduction Complex microparts with many small features generally exhibit a physical effect through the sludge model, which needs to be scanned into a point cloud in industrial production. The threedimensional (3D) models of these complex microparts are reconstructed using reverse
* Xiaogang Ji [email protected]
1
School of Mechanical Engineering, Jiangnan University, Wuxi 214122 Jiangsu, China
2
Jiangsu Key Laboratory of Advanced Food Manufacturing Equipment and Technology (Jiangnan University), Wuxi 214122 Jiangsu, China
Multimedia Tools and Applications
engineering. Point cloud segmentation, which is based on features and functions, is the key to reconstruct the models. Each segmented area with the same properties [14, 20] corresponds to a basic shape on the model. The manual segmentation method involves a point-by-point evaluation of a region. When using this method, the computational load is heavy, the efficiency is low; and the boundary of the segmentation is not fair. The intelligent segmentation algorithm, which is based on geometric quantities such as curvature and normal vectors, has become a hot research topic owing to its small dependence on labor. The algorithm can be classified mainly into edge-based and surface-based methods. The edge-based method [8, 9, 13, 18] extracts points with sudden changes in differential geometric quantities such as feature points and then connects
Data Loading...
