class mrpt::maps::CMultiMetricMap

Overview

This class stores any customizable set of metric maps.

The internal metric maps can be accessed directly by the user as smart pointers with CMultiMetricMap::mapByIndex() or via iterator s. The utility of this container is to operate on several maps simultaneously: update them by inserting observations, evaluate the likelihood of one observation by fusing (multiplying) the likelihoods over the different maps, etc.

These kinds of metric maps can be kept inside (list may be incomplete, refer to classes derived from mrpt::maps::CMetricMap):

mrpt::maps::CSimplePointsMap : For 2D or 3D range scans, …
mrpt::maps::COccupancyGridMap2D : 2D, horizontal laser range scans, at different altitudes.
mrpt::maps::COccupancyGridMap3D : 3D occupancy voxel map.
mrpt::maps::COctoMap : For 3D occupancy grids of variable resolution, with octrees (based on the library octomap).

mrpt::maps::CVoxelMap or mrpt::maps::CVoxelMapRGB : 3D sparse voxel maps.

mrpt::maps::CColouredOctoMap : The same than above, but nodes can store RGB data appart from occupancy.
mrpt::maps::CLandmarksMap : For visual landmarks,etc…
mrpt::maps::CGasConcentrationGridMap2D : For gas concentration maps.
mrpt::maps::CWirelessPowerGridMap2D : For wifi power maps.
mrpt::maps::CBeaconMap : For range-only SLAM.
mrpt::maps::CHeightGridMap2D : For elevation maps of height for each (x,y) location (Digital elevation model, DEM)
mrpt::maps::CHeightGridMap2D_MRF : DEMs as Markov Random Field (MRF)
mrpt::maps::CReflectivityGridMap2D : For maps of “reflectivity” for each (x,y) location.
mrpt::maps::CColouredPointsMap : For point map with color.
mrpt::maps::CWeightedPointsMap : For point map with weights (capable of “fusing”).

See CMultiMetricMap::setListOfMaps() for the method for initializing this class programmatically. See also TSetOfMetricMapInitializers::loadFromConfigFile for a template of “.ini”-like configuration file that can be used to define which maps to create and all their parameters. Alternatively, the list of maps is public so it can be directly manipulated/accessed in CMultiMetricMap::maps

#include <mrpt/maps/CMultiMetricMap.h>

class CMultiMetricMap: public mrpt::maps::CMetricMap
{
public:
    // typedefs

    typedef std::deque<mrpt::maps::CMetricMap::Ptr> TListMaps;
    typedef TListMaps::iterator iterator;
    typedef TListMaps::const_iterator const_iterator;

    // fields

    TListMaps maps;

    // construction

    CMultiMetricMap();
    CMultiMetricMap(const TSetOfMetricMapInitializers& initializers);
    CMultiMetricMap(const CMultiMetricMap& o);
    CMultiMetricMap(CMultiMetricMap&&);

    // methods

    iterator begin();
    const_iterator begin() const;
    iterator end();
    const_iterator end() const;
    mrpt::maps::CMetricMap::Ptr mapByIndex(size_t idx) const;

    template <typename T>
    T::Ptr mapByClass(size_t ith = 0) const;

    template <typename T>
    std::size_t countMapsByClass() const;

    CMultiMetricMap& operator = (const CMultiMetricMap& o);
    CMultiMetricMap& operator = (CMultiMetricMap&&);
    void setListOfMaps(const mrpt::maps::TSetOfMetricMapInitializers& init);
    virtual bool isEmpty() const;

    virtual void determineMatching2D(
        const mrpt::maps::CMetricMap* otherMap,
        const mrpt::poses::CPose2D& otherMapPose,
        mrpt::tfest::TMatchingPairList& correspondences,
        const mrpt::maps::TMatchingParams& params,
        mrpt::maps::TMatchingExtraResults& extraResults
        ) const;

    virtual float compute3DMatchingRatio(const mrpt::maps::CMetricMap* otherMap, const mrpt::poses::CPose3D& otherMapPose, const TMatchingRatioParams& params) const;
    virtual void saveMetricMapRepresentationToFile(const std::string& filNamePrefix) const;
    virtual void auxParticleFilterCleanUp();
    virtual void getVisualizationInto(mrpt::opengl::CSetOfObjects& o) const;
    virtual const mrpt::maps::CSimplePointsMap* getAsSimplePointsMap() const;
    virtual std::string asString() const;
};

Inherited Members

public:
    // methods

    virtual bool isEmpty() const = 0;
    virtual void saveMetricMapRepresentationToFile(const std::string& filNamePrefix) const = 0;
    virtual std::string asString() const = 0;
    virtual void getVisualizationInto(mrpt::opengl::CSetOfObjects& o) const = 0;

Fields

TListMaps maps

The list of metric maps in this object.

Use dynamic_cast or smart pointer-based downcast to access maps by their actual type. You can directly manipulate this list. Helper methods to initialize it are described in the docs of CMultiMetricMap

Construction

CMultiMetricMap()

Default ctor: empty list of maps.

CMultiMetricMap(const TSetOfMetricMapInitializers& initializers)

Constructor with a list of map initializers.

Parameters:

initializers

One internal map will be created for each entry in this “TSetOfMetricMapInitializers” struct.

CMultiMetricMap(const CMultiMetricMap& o)

Creates a deep copy.

CMultiMetricMap(CMultiMetricMap&&)

Move ctor.

Methods

mrpt::maps::CMetricMap::Ptr mapByIndex(size_t idx) const

Gets the i-th map.

Parameters:

std::runtime_error

On out-of-bounds

template <typename T>
T::Ptr mapByClass(size_t ith = 0) const

Returns the i’th map of a given class (or of a derived class), or empty smart pointer if there is no such map.

Example:

 COccupancyGridMap2D::Ptr obs =
multimap.mapByClass<COccupancyGridMap2D>();

By default (ith=0), the first match is returned.

template <typename T>
std::size_t countMapsByClass() const

Count how many maps exist of the given class (or derived class)

CMultiMetricMap& operator = (const CMultiMetricMap& o)

Creates a deep copy.

CMultiMetricMap& operator = (CMultiMetricMap&&)

Move operator.

void setListOfMaps(const mrpt::maps::TSetOfMetricMapInitializers& init)

Sets the list of internal map according to the passed list of map initializers (current maps will be deleted)

virtual bool isEmpty() const

Returns true if all maps returns true in their isEmpty() method.

virtual void determineMatching2D(
    const mrpt::maps::CMetricMap* otherMap,
    const mrpt::poses::CPose2D& otherMapPose,
    mrpt::tfest::TMatchingPairList& correspondences,
    const mrpt::maps::TMatchingParams& params,
    mrpt::maps::TMatchingExtraResults& extraResults
    ) const

Computes the matching between this and another 2D point map, which includes finding:

The set of points pairs in each map
The mean squared distance between corresponding pairs.

The algorithm is:

For each point in “otherMap”:
- Transform the point according to otherMapPose
- Search with a KD-TREE the closest correspondences in “this” map.
- Add to the set of candidate matchings, if it passes all the thresholds in params.

This method is the most time critical one into ICP-like algorithms.

Parameters:

otherMap	[IN] The other map to compute the matching with.
otherMapPose	[IN] The pose of the other map as seen from “this”.
params	[IN] Parameters for the determination of pairings.
correspondences	[OUT] The detected matchings pairs.
extraResults	[OUT] Other results.