点云配准方法.ppt

* 1. Point Cloud Registration With Target Control Error equation: 1. Point Cloud Registration With Target Control * New approximate value: Iteration: Finally, we get the values of parameter: 2. Iterative Closest Point(ICP) Algorithm * Basic thought: Firstly, from a point set, a bar curve, or a surface to find the closest point that corresponds to one point, and then uses this result to find two corresponding point sets. Finally, we find out the corresponding point set and the corresponding transformation matrix. Basic steps: Step1:Search for the nearest points Step2:Solve transformation relations Step3:Application transformation Step4:Repeated iteration 3 Quaternion Method What is Quaternion ? * Unit quaternions Rotation matrix R 3. Quaternion Method Definite translation vector: * Complete registration station: Assume that: Minimum objective function: The centroid of the set of basic points : 3. Quaternion Method * Orthogonal covariance matrix of point set: Antisymmetric matrix’s cyclic components: Construct a symmetric matrix: 3. Quaternion Method * Translation matrix: The unit eigenvector corresponding to the maximum eigenvalue of matrix : Conclusions Thank you！ * * 数字城市是真实城市环境与虚拟城市环境的融合与协同，三维GIS在其中发挥重要的时空信息承载和纽带作用，并提供了有力的可视化分析方法和支撑手段，正日益成为政府、企业和公众不可或缺的科学的、定量与定性结合的、直观方便的助手。 地形的实现方式一：四叉树LOD方式，会产生裂缝 地形的实现方式二：CLOD，不会产生裂缝 * 仪器科学与工程学院 仪器科学与工程学院 A Brief Introduction Of Point Cloud Registration Method * Point Cloud Registration Method Point Cloud Registration With Target Control 3 4 Quaternion Method 2 1 Iterative Closest Point(ICP) Algorithm 4 Outline Two point cloud sets of A and B, are conformations of the same points in different coordinate systems Rigid transformation: Objective function: 1. Point Cloud Registration With Target Control * 面 （a）Leica standard rotation plane target （b） Mensi plane target Figure 1 Different kinds of plane target （d） Mensi