Template Struct KDTreeEigenMatrixAdaptor

Defined in File nanoflann.hpp

Struct Documentation

template<class MatrixType, int32_t DIM = -1, class Distance = nanoflann::metric_L2, bool row_major = true> struct KDTreeEigenMatrixAdaptor

An L2-metric KD-tree adaptor for working with data directly stored in an Eigen Matrix, without duplicating the data storage. You can select whether a row or column in the matrix represents a point in the state space.

Example of usage:

Eigen::Matrix<num_t,Eigen::Dynamic,Eigen::Dynamic>  mat;

// Fill out "mat"...
using my_kd_tree_t = nanoflann::KDTreeEigenMatrixAdaptor<
  Eigen::Matrix<num_t,Dynamic,Dynamic>>;

const int max_leaf = 10;
my_kd_tree_t mat_index(mat, max_leaf);
mat_index.index->...

Template Parameters:

DIM – If set to >0, it specifies a compile-time fixed dimensionality for the points in the data set, allowing more compiler optimizations.
Distance – The distance metric to use: nanoflann::metric_L1, nanoflann::metric_L2, nanoflann::metric_L2_Simple, etc.
row_major – If set to true the rows of the matrix are used as the points, if set to false the columns of the matrix are used as the points.

Interface expected by KDTreeSingleIndexAdaptor

inline const self_t &derived() const

inline self_t &derived()

inline Size kdtree_get_point_count() const

inline num_t kdtree_get_pt(const IndexType idx, size_t dim) const

template<class BBOX> inline bool kdtree_get_bbox(BBOX&) const

Public Types

using self_t = KDTreeEigenMatrixAdaptor<MatrixType, DIM, Distance, row_major>

using num_t = typename MatrixType::Scalar

using IndexType = typename MatrixType::Index

using metric_t = typename Distance::template traits<num_t, self_t, IndexType>::distance_t

using index_t = KDTreeSingleIndexAdaptor<metric_t, self_t, row_major ? MatrixType::ColsAtCompileTime : MatrixType::RowsAtCompileTime, IndexType>

using Offset = typename index_t::Offset: The kd-tree index for the user to call its methods as usual with any other FLANN index.

using Size = typename index_t::Size

using Dimension = typename index_t::Dimension

Public Functions

inline explicit KDTreeEigenMatrixAdaptor(const Dimension dimensionality, const std::reference_wrapper<const MatrixType> &mat, const int leaf_max_size = 10): Constructor: takes a const ref to the matrix object with the data points.

KDTreeEigenMatrixAdaptor(const self_t&) = delete: Deleted copy constructor

inline ~KDTreeEigenMatrixAdaptor()

inline void query(const num_t *query_point, const Size num_closest, IndexType *out_indices, num_t *out_distances) const: Query for the num_closest closest points to a given point (entered as query_point[0:dim-1]). Note that this is a short-cut method for index->findNeighbors(). The user can also call index->… methods as desired.

Note

If L2 norms are used, all returned distances are actually squared distances.

Public Members

index_t *index_

const std::reference_wrapper<const MatrixType> m_data_matrix