首页 > 代码库 > ICP(IterativeClosestPoint)

ICP(IterativeClosestPoint)

含义:

 

  • IterativeClosestPointNonLinear is an ICP variant that uses Levenberg-Marquardt optimization backend.

 

The resultant transformation is optimized as a quaternion.

The algorithm has several termination criteria:(终止条件)

    1. Number of iterations has reached the maximum user imposed number of iterations (via setMaximumIterations)
    2. The epsilon (difference) between the previous transformation and the current estimated transformation is smaller than an user imposed value (via setTransformationEpsilon)
    3. The sum of Euclidean squared errors is smaller than a user defined threshold (via setEuclideanFitnessEpsilon)
      来源: http://docs.pointclouds.org/trunk/classpcl_1_1_iterative_closest_point_non_linear.html#details
  • ICP优化步骤,既可以用于局部几何特征不变描述算子中也可以用于全局几何特征不变描述算子。

文献:

  1. A method for registration for 3-D shapes.pdf
 

用法:

 

  • 是一种匹配算法,较常用于配准。主要是用来寻找一个点云和另个点云的最佳变换关系,一般是点和点的匹配,让两个点云近似的去重合,它会不断的对齐点云去改进变换的效果,直到达到终止条件(最大迭代次数,距离或是误差范围)。
  • ICP通过迭代去更新结果,效果不错,但常常比较耗时。
  • ICP的各种改进算法,常见的ICP算法,如:
    • 点一线ICP匹配、
    • 线一线ICP匹配、
    • 线一面ICP匹配等。

 

 

 

参考:

  1. 基于ROS和点云库的室内三维物体识别与姿态估计_鄢武



null



ICP(IterativeClosestPoint)