研究目的
To introduce an improved method to evaluate the quality of registration methods for mobile laser scanning (MLS) point clouds, based on overcoming shortfalls in existing methods.
研究成果
The proposed error metric is more suitable for evaluating and comparing MLS point cloud registration outputs than state-of-the-art metrics. LSPFA outperforms point-based matching techniques for registering MLS point clouds.
研究不足
The paper does not explicitly mention limitations, but the experimental datasets are sparse and heterogeneous, with an average point spacing of 7 cm and an average profile spacing of 25 cm, which could affect the registration accuracy.
1：Experimental Design and Method Selection:
The paper reviews existing target-free matching techniques and introduces a new error metric for evaluating the quality of point cloud registration.
2：Sample Selection and Data Sources:
Real datasets captured near the campus of Curtin University using a Dynascan MDL S250 system were used.
3：List of Experimental Equipment and Materials:
A Dynascan MDL S250 system was used for capturing MLS point clouds.
4：Experimental Procedures and Operational Workflow:
The paper compares point-based matching (both point-to-point and point-to-plane) and plane-based matching (LSPFA) in MLS point cloud registration.
5：Data Analysis Methods:
The proposed error metric utilises the RMS values of points fitted onto check planes, taking into account the orientations of the check planes.
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
