研究目的
To propose a subtle feature extraction algorithm for laser scanning measurement of large-scale irregular surfaces in reverse engineering, addressing the challenges of large point cloud data and complex feature recognition.
研究成果
The proposed subtle feature extraction algorithm effectively identifies and locates subtle features on large-scale irregular surfaces, outperforming traditional algorithms in accuracy and robustness. It provides a theoretical basis for point cloud feature extraction in reverse engineering applications.
研究不足
The algorithm may still experience some data loss on ring weld information, requiring manual adjustment for subsequent processing. The accuracy of feature recognition needs further improvement.
1：Experimental Design and Method Selection:
The study involves calculating the L1-median point as the neighborhood center, introducing k+1 neighbors for feature description, and using a multi-threshold Poisson region growth algorithm for feature extraction.
2：Sample Selection and Data Sources:
Point cloud data from a large semi-ellipsoid crown workpiece obtained via a laser scanning measurement system.
3：List of Experimental Equipment and Materials:
KEYENCE LR-ZB250AN laser scanning sensor with a measuring accuracy of 16um.
4：Experimental Procedures and Operational Workflow:
Calibration of the laser sensor, manual movement of the measuring arm, and equally spaced data acquisition on the workpiece driven by a rotary table.
5：Data Analysis Methods:
Comparison of the proposed algorithm with traditional algorithms (ISS algorithm by Zhong and algorithm by Pauly) in terms of feature extraction effectiveness and accuracy.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
