Template Class KDTreeSingleIndexDynamicAdaptor_

Defined in File nanoflann.hpp

Inheritance Relationships

Base Type

public nanoflann::KDTreeBaseClass< KDTreeSingleIndexDynamicAdaptor_< Distance, DatasetAdaptor, -1, uint32_t >, Distance, DatasetAdaptor, -1, uint32_t > (Template Class KDTreeBaseClass)

Class Documentation

template<typename Distance, class DatasetAdaptor, int32_t DIM = -1, typename IndexType = uint32_t> class KDTreeSingleIndexDynamicAdaptor_ : public nanoflann::KDTreeBaseClass<KDTreeSingleIndexDynamicAdaptor_<Distance, DatasetAdaptor, -1, uint32_t>, Distance, DatasetAdaptor, -1, uint32_t>

kd-tree dynamic index

Contains the k-d trees and other information for indexing a set of points for nearest-neighbor matching.

The class “DatasetAdaptor” must provide the following interface (can be non-virtual, inlined methods):

  // Must return the number of data poins
  size_t kdtree_get_point_count() const { ... }

  // Must return the dim'th component of the idx'th point in the class:
  T kdtree_get_pt(const size_t idx, const size_t dim) const { ... }

  // Optional bounding-box computation: return false to default to a standard
bbox computation loop.
  //   Return true if the BBOX was already computed by the class and returned
in "bb" so it can be avoided to redo it again.
  //   Look at bb.size() to find out the expected dimensionality (e.g. 2 or 3
for point clouds) template <class BBOX> bool kdtree_get_bbox(BBOX &bb) const
  {
     bb[0].low = ...; bb[0].high = ...;  // 0th dimension limits
     bb[1].low = ...; bb[1].high = ...;  // 1st dimension limits
     ...
     return true;
  }

Template Parameters:

DatasetAdaptor – The user-provided adaptor (see comments above).
Distance – The distance metric to use: nanoflann::metric_L1, nanoflann::metric_L2, nanoflann::metric_L2_Simple, etc.
DIM – Dimensionality of data points (e.g. 3 for 3D points)
IndexType – Type of the arguments with which the data can be accessed (e.g. float, double, int64_t, T*)

Query methods

template<typename RESULTSET> inline bool findNeighbors(RESULTSET &result, const ElementType *vec, const SearchParameters &searchParams = {}) const

Find set of nearest neighbors to vec[0:dim-1]. Their indices are stored inside the result object. This is the core search function, all others are wrappers around this one.

Note

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

Parameters:

result – The result object in which the indices of the nearest-neighbors are stored.
vec – The vector of the query point for which to search the nearest neighbors.
searchParams – Optional parameters for the search.

Template Parameters:

RESULTSET – Should be any ResultSet<DistanceType>

Returns:

True if the requested neighbors could be found.

inline Size knnSearch(const ElementType *query_point, const Size num_closest, IndexType *out_indices, DistanceType *out_distances, const SearchParameters &searchParams = {}) const

Find the “num_closest” nearest neighbors to the query_point[0:dim-1]. Their indices are stored inside the result object.

See also

radiusSearch, findNeighbors

Note

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

Note

Only the first N entries in out_indices and out_distances will be valid. Return may be less than num_closest only if the number of elements in the tree is less than num_closest.

Returns:: Number N of valid points in the result set.

inline Size radiusSearch(const ElementType *query_point, const DistanceType &radius, std::vector<ResultItem<IndexType, DistanceType>> &IndicesDists, const SearchParameters &searchParams = {}) const

Find all the neighbors to query_point[0:dim-1] within a maximum radius. The output is given as a vector of pairs, of which the first element is a point index and the second the corresponding distance. Previous contents of IndicesDists are cleared.

If searchParams.sorted==true, the output list is sorted by ascending distances.

For a better performance, it is advisable to do a .reserve() on the vector if you have any wild guess about the number of expected matches.

Note

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

Returns:: The number of points within the given radius (i.e. indices.size() or dists.size() )

template<class SEARCH_CALLBACK> inline Size radiusSearchCustomCallback(const ElementType *query_point, SEARCH_CALLBACK &resultSet, const SearchParameters &searchParams = {}) const

Just like radiusSearch() but with a custom callback class for each point found in the radius of the query. See the source of RadiusResultSet<> as a start point for your own classes.

See also

radiusSearch

Public Types

using Base = typename nanoflann::KDTreeBaseClass<nanoflann::KDTreeSingleIndexDynamicAdaptor_<Distance, DatasetAdaptor, DIM, IndexType>, Distance, DatasetAdaptor, DIM, IndexType>

using ElementType = typename Base::ElementType

using DistanceType = typename Base::DistanceType

using Offset = typename Base::Offset

using Size = typename Base::Size

using Dimension = typename Base::Dimension

using Node = typename Base::Node

using NodePtr = Node*

using Interval = typename Base::Interval

using BoundingBox = typename Base::BoundingBox: Define “BoundingBox” as a fixed-size or variable-size container depending on “DIM”

using distance_vector_t = typename Base::distance_vector_t: Define “distance_vector_t” as a fixed-size or variable-size container depending on “DIM”

Public Functions

inline KDTreeSingleIndexDynamicAdaptor_(const Dimension dimensionality, const DatasetAdaptor &inputData, std::vector<int> &treeIndex, const KDTreeSingleIndexAdaptorParams &params = KDTreeSingleIndexAdaptorParams())

KDTree constructor

Refer to docs in README.md or online in https://github.com/jlblancoc/nanoflann

The KD-Tree point dimension (the length of each point in the datase, e.g. 3 for 3D points) is determined by means of:

The DIM template parameter if >0 (highest priority)
Otherwise, the dimensionality parameter of this constructor.

Parameters:

inputData – Dataset with the input features. Its lifetime must be equal or longer than that of the instance of this class.
params – Basically, the maximum leaf node size

KDTreeSingleIndexDynamicAdaptor_(const KDTreeSingleIndexDynamicAdaptor_ &rhs) = default: Explicitly default the copy constructor

inline KDTreeSingleIndexDynamicAdaptor_ operator=(const KDTreeSingleIndexDynamicAdaptor_ &rhs): Assignment operator definiton

inline void buildIndex(): Builds the index

inline void computeBoundingBox(BoundingBox &bbox)

template<class RESULTSET> inline void searchLevel(RESULTSET &result_set, const ElementType *vec, const NodePtr node, DistanceType mindist, distance_vector_t &dists, const float epsError) const

Performs an exact search in the tree starting from a node.

Template Parameters:: RESULTSET – Should be any ResultSet<DistanceType>

inline void saveIndex(std::ostream &stream)

Stores the index in a binary file. IMPORTANT NOTE: The set of data points is NOT stored in the file, so when loading the index object it must be constructed associated to the same source of data points used while building it. See the example: examples/saveload_example.cpp

See also

loadIndex

inline void loadIndex(std::istream &stream)

Loads a previous index from a binary file. IMPORTANT NOTE: The set of data points is NOT stored in the file, so the index object must be constructed associated to the same source of data points used while building the index. See the example: examples/saveload_example.cpp

See also

loadIndex

Public Members

const DatasetAdaptor &dataset_

The source of our data.

The dataset used by this index

KDTreeSingleIndexAdaptorParams index_params_

std::vector<int> &treeIndex_

Distance distance_