Preliminary Point Cloud Data Analysis for Thin-Walled Part Defect Detection based on Point Cloud Data Processing
This paper studies the processing of point cloud data for objects, focusing on improving the bilateral filtering algorithm to address poor denoising effects by removing edge noise while preserving features. It employs voxel grid downsampling to streamline large datasets. During the point cloud registration phase, a standard plane is first fitted before registration, introducing the FPFH and SAC coarse registration and ICP fine registration algorithms to enhance registration outcomes. In the defect detection task, defect data points are extracted by calculating Euclidean distances, and the greedy projection triangulation method is used to reconstruct surfaces for visualization and quantification. This approach provides data support for subsequent work, aiding in assessing the extent of object damage.
References
- Wu Lushen, Li Ze, Chen Huawai, et al. Research on Improved Resampling Algorithm. Journal of Mechanical Design and Manufacturing, 2015(04): 244-247.
- Yuan Hua, Pang Jiankang, Mo Jianwen. Research on Point Cloud Simplification Algorithm Based on Voxel Grid Downsampling. Television Technology, 2015, 39(17): 43-47.
- Rusu R B, Blodow N, Marton Z C, et al. Aligning point cloud views using persistent feature histograms[C]//2008 IEEE/RSJ international conference on intelligent robots and systems. IEEE, 2008: 3384-3391.
- Rusu R B, Blodow N, Beetz M. Fast point feature histograms (FPFH) for 3D registration[C]//2009 IEEE international conference on robotics and automation. IEEE, 2009: 3212-3217.
- Huang Y, Da F, Tao H. An automatic registration algorithm for point cloud based on feature extraction[J]. Chin. J. Lasers, 2015, 42(3): 242-248.