The key point is also called point of interest, it is the stability point set, distinguishing the 2D image or a 3D point cloud model or a curved surface, can be obtained by defining testing standards, technically speaking, the number of key points with reducing the amount of data than the original image or point cloud many sub combined together to form a critical point descriptors used to form the original data representation, and yet descriptive representation and the local feature descriptor, which speeds up the subsequent identifying, tracing the data processing speed, therefore, become the key technology in 2D and 3D critical information processing technologies.
1. The key point extraction steps:
virtual void setSearchSurface (const PointCloudInConstPtr &cloud)
设置搜索时所用搜索点云， cloud 为指向点云对象的指针引用．
void setSearchMethod (const KdTreePtr &tree)
设置内部算法实现时所用的搜索对象， tree 为指向 kdtree 或 octree 对应的指针引用．
void setKSearch (int k)
设置 k 近邻搜索时所用的 k 参数。
void setRadiusSearch (double radius)
设置半径搜索时所用的半径 radius 参数。
int searchForNeighbors (int index, double parameter, std::vector< int > &indices, std::vector< float > &distances) const
采用 setSearchMethod 设置的搜索对象 ，以及 setSearchSurface 设置的搜索点云 ，进行近邻搜索，返回近邻在点云中的索引向盘 indices ，以及对应的距离向盘 distanee ，其中 index 为查询点的索引， parameter 为搜索时所用的参数半径 radius 或者 k.