OpenCV  4.9.0-dev
Open Source Computer Vision
Loading...
Searching...
No Matches
Classes | Public Types | Public Member Functions | Static Public Member Functions | Protected Member Functions | Static Protected Member Functions | Protected Attributes | List of all members
cv::DescriptorMatcher Class Referenceabstract

Abstract base class for matching keypoint descriptors. More...

#include <opencv2/features2d.hpp>

Inheritance diagram for cv::DescriptorMatcher:
Collaboration diagram for cv::DescriptorMatcher:

Classes

class  DescriptorCollection
 

Public Types

enum  MatcherType {
  FLANNBASED = 1 ,
  BRUTEFORCE = 2 ,
  BRUTEFORCE_L1 = 3 ,
  BRUTEFORCE_HAMMING = 4 ,
  BRUTEFORCE_HAMMINGLUT = 5 ,
  BRUTEFORCE_SL2 = 6
}
 

Public Member Functions

virtual ~DescriptorMatcher ()
 
virtual void add (InputArrayOfArrays descriptors)
 Adds descriptors to train a CPU(trainDescCollectionis) or GPU(utrainDescCollectionis) descriptor collection.
 
virtual void clear () CV_OVERRIDE
 Clears the train descriptor collections.
 
virtual CV_NODISCARD_STD Ptr< DescriptorMatcherclone (bool emptyTrainData=false) const =0
 Clones the matcher.
 
virtual bool empty () const CV_OVERRIDE
 Returns true if there are no train descriptors in the both collections.
 
const std::vector< Mat > & getTrainDescriptors () const
 Returns a constant link to the train descriptor collection trainDescCollection .
 
virtual bool isMaskSupported () const =0
 Returns true if the descriptor matcher supports masking permissible matches.
 
void knnMatch (InputArray queryDescriptors, InputArray trainDescriptors, std::vector< std::vector< DMatch > > &matches, int k, InputArray mask=noArray(), bool compactResult=false) const
 Finds the k best matches for each descriptor from a query set.
 
void knnMatch (InputArray queryDescriptors, std::vector< std::vector< DMatch > > &matches, int k, InputArrayOfArrays masks=noArray(), bool compactResult=false)
 
void match (InputArray queryDescriptors, InputArray trainDescriptors, std::vector< DMatch > &matches, InputArray mask=noArray()) const
 Finds the best match for each descriptor from a query set.
 
void match (InputArray queryDescriptors, std::vector< DMatch > &matches, InputArrayOfArrays masks=noArray())
 
void radiusMatch (InputArray queryDescriptors, InputArray trainDescriptors, std::vector< std::vector< DMatch > > &matches, float maxDistance, InputArray mask=noArray(), bool compactResult=false) const
 For each query descriptor, finds the training descriptors not farther than the specified distance.
 
void radiusMatch (InputArray queryDescriptors, std::vector< std::vector< DMatch > > &matches, float maxDistance, InputArrayOfArrays masks=noArray(), bool compactResult=false)
 
virtual void read (const FileNode &) CV_OVERRIDE
 Reads algorithm parameters from a file storage.
 
void read (const String &fileName)
 
virtual void train ()
 Trains a descriptor matcher.
 
void write (const Ptr< FileStorage > &fs, const String &name) const
 
void write (const String &fileName) const
 
virtual void write (FileStorage &) const CV_OVERRIDE
 Stores algorithm parameters in a file storage.
 
void write (FileStorage &fs, const String &name) const
 
- Public Member Functions inherited from cv::Algorithm
 Algorithm ()
 
virtual ~Algorithm ()
 
virtual String getDefaultName () const
 
virtual void save (const String &filename) const
 
void write (const Ptr< FileStorage > &fs, const String &name=String()) const
 
void write (FileStorage &fs, const String &name) const
 

Static Public Member Functions

static Ptr< DescriptorMatchercreate (const DescriptorMatcher::MatcherType &matcherType)
 
static Ptr< DescriptorMatchercreate (const String &descriptorMatcherType)
 Creates a descriptor matcher of a given type with the default parameters (using default constructor).
 
- Static Public Member Functions inherited from cv::Algorithm
template<typename _Tp >
static Ptr< _Tpload (const String &filename, const String &objname=String())
 Loads algorithm from the file.
 
template<typename _Tp >
static Ptr< _TploadFromString (const String &strModel, const String &objname=String())
 Loads algorithm from a String.
 
template<typename _Tp >
static Ptr< _Tpread (const FileNode &fn)
 Reads algorithm from the file node.
 

Protected Member Functions

void checkMasks (InputArrayOfArrays masks, int queryDescriptorsCount) const
 
virtual void knnMatchImpl (InputArray queryDescriptors, std::vector< std::vector< DMatch > > &matches, int k, InputArrayOfArrays masks=noArray(), bool compactResult=false)=0
 
virtual void radiusMatchImpl (InputArray queryDescriptors, std::vector< std::vector< DMatch > > &matches, float maxDistance, InputArrayOfArrays masks=noArray(), bool compactResult=false)=0
 
- Protected Member Functions inherited from cv::Algorithm
void writeFormat (FileStorage &fs) const
 

Static Protected Member Functions

static CV_NODISCARD_STD Mat clone_op (Mat m)
 
static bool isMaskedOut (InputArrayOfArrays masks, int queryIdx)
 
static bool isPossibleMatch (InputArray mask, int queryIdx, int trainIdx)
 

Protected Attributes

std::vector< MattrainDescCollection
 Collection of descriptors from train images.
 
std::vector< UMatutrainDescCollection
 

Detailed Description

Abstract base class for matching keypoint descriptors.

It has two groups of match methods: for matching descriptors of an image with another image or with an image set.

Member Enumeration Documentation

◆ MatcherType

Enumerator
FLANNBASED 
BRUTEFORCE 
BRUTEFORCE_L1 
BRUTEFORCE_HAMMING 
BRUTEFORCE_HAMMINGLUT 
BRUTEFORCE_SL2 

Constructor & Destructor Documentation

◆ ~DescriptorMatcher()

virtual cv::DescriptorMatcher::~DescriptorMatcher ( )
virtual

Member Function Documentation

◆ add()

virtual void cv::DescriptorMatcher::add ( InputArrayOfArrays  descriptors)
virtual
Python:
cv.DescriptorMatcher.add(descriptors) -> None

Adds descriptors to train a CPU(trainDescCollectionis) or GPU(utrainDescCollectionis) descriptor collection.

If the collection is not empty, the new descriptors are added to existing train descriptors.

Parameters
descriptorsDescriptors to add. Each descriptors[i] is a set of descriptors from the same train image.

Reimplemented in cv::FlannBasedMatcher.

◆ checkMasks()

void cv::DescriptorMatcher::checkMasks ( InputArrayOfArrays  masks,
int  queryDescriptorsCount 
) const
protected

◆ clear()

virtual void cv::DescriptorMatcher::clear ( )
virtual
Python:
cv.DescriptorMatcher.clear() -> None

Clears the train descriptor collections.

Reimplemented from cv::Algorithm.

Reimplemented in cv::FlannBasedMatcher.

◆ clone()

virtual CV_NODISCARD_STD Ptr< DescriptorMatcher > cv::DescriptorMatcher::clone ( bool  emptyTrainData = false) const
pure virtual
Python:
cv.DescriptorMatcher.clone([, emptyTrainData]) -> retval

Clones the matcher.

Parameters
emptyTrainDataIf emptyTrainData is false, the method creates a deep copy of the object, that is, copies both parameters and train data. If emptyTrainData is true, the method creates an object copy with the current parameters but with empty train data.

Implemented in cv::BFMatcher, and cv::FlannBasedMatcher.

◆ clone_op()

static CV_NODISCARD_STD Mat cv::DescriptorMatcher::clone_op ( Mat  m)
inlinestaticprotected
Here is the call graph for this function:

◆ create() [1/2]

static Ptr< DescriptorMatcher > cv::DescriptorMatcher::create ( const DescriptorMatcher::MatcherType matcherType)
static
Python:
cv.DescriptorMatcher.create(descriptorMatcherType) -> retval
cv.DescriptorMatcher.create(matcherType) -> retval
cv.DescriptorMatcher_create(descriptorMatcherType) -> retval
cv.DescriptorMatcher_create(matcherType) -> retval

◆ create() [2/2]

static Ptr< DescriptorMatcher > cv::DescriptorMatcher::create ( const String descriptorMatcherType)
static
Python:
cv.DescriptorMatcher.create(descriptorMatcherType) -> retval
cv.DescriptorMatcher.create(matcherType) -> retval
cv.DescriptorMatcher_create(descriptorMatcherType) -> retval
cv.DescriptorMatcher_create(matcherType) -> retval

Creates a descriptor matcher of a given type with the default parameters (using default constructor).

Parameters
descriptorMatcherTypeDescriptor matcher type. Now the following matcher types are supported:
  • BruteForce (it uses L2 )
  • BruteForce-L1
  • BruteForce-Hamming
  • BruteForce-Hamming(2)
  • FlannBased

◆ empty()

virtual bool cv::DescriptorMatcher::empty ( ) const
virtual
Python:
cv.DescriptorMatcher.empty() -> retval

Returns true if there are no train descriptors in the both collections.

Reimplemented from cv::Algorithm.

◆ getTrainDescriptors()

const std::vector< Mat > & cv::DescriptorMatcher::getTrainDescriptors ( ) const
Python:
cv.DescriptorMatcher.getTrainDescriptors() -> retval

Returns a constant link to the train descriptor collection trainDescCollection .

◆ isMaskedOut()

static bool cv::DescriptorMatcher::isMaskedOut ( InputArrayOfArrays  masks,
int  queryIdx 
)
staticprotected

◆ isMaskSupported()

virtual bool cv::DescriptorMatcher::isMaskSupported ( ) const
pure virtual
Python:
cv.DescriptorMatcher.isMaskSupported() -> retval

Returns true if the descriptor matcher supports masking permissible matches.

Implemented in cv::BFMatcher, and cv::FlannBasedMatcher.

◆ isPossibleMatch()

static bool cv::DescriptorMatcher::isPossibleMatch ( InputArray  mask,
int  queryIdx,
int  trainIdx 
)
staticprotected

◆ knnMatch() [1/2]

void cv::DescriptorMatcher::knnMatch ( InputArray  queryDescriptors,
InputArray  trainDescriptors,
std::vector< std::vector< DMatch > > &  matches,
int  k,
InputArray  mask = noArray(),
bool  compactResult = false 
) const
Python:
cv.DescriptorMatcher.knnMatch(queryDescriptors, trainDescriptors, k[, mask[, compactResult]]) -> matches
cv.DescriptorMatcher.knnMatch(queryDescriptors, k[, masks[, compactResult]]) -> matches

Finds the k best matches for each descriptor from a query set.

Parameters
queryDescriptorsQuery set of descriptors.
trainDescriptorsTrain set of descriptors. This set is not added to the train descriptors collection stored in the class object.
maskMask specifying permissible matches between an input query and train matrices of descriptors.
matchesMatches. Each matches[i] is k or less matches for the same query descriptor.
kCount of best matches found per each query descriptor or less if a query descriptor has less than k possible matches in total.
compactResultParameter used when the mask (or masks) is not empty. If compactResult is false, the matches vector has the same size as queryDescriptors rows. If compactResult is true, the matches vector does not contain matches for fully masked-out query descriptors.

These extended variants of DescriptorMatcher::match methods find several best matches for each query descriptor. The matches are returned in the distance increasing order. See DescriptorMatcher::match for the details about query and train descriptors.

◆ knnMatch() [2/2]

void cv::DescriptorMatcher::knnMatch ( InputArray  queryDescriptors,
std::vector< std::vector< DMatch > > &  matches,
int  k,
InputArrayOfArrays  masks = noArray(),
bool  compactResult = false 
)
Python:
cv.DescriptorMatcher.knnMatch(queryDescriptors, trainDescriptors, k[, mask[, compactResult]]) -> matches
cv.DescriptorMatcher.knnMatch(queryDescriptors, k[, masks[, compactResult]]) -> matches

This is an overloaded member function, provided for convenience. It differs from the above function only in what argument(s) it accepts.

Parameters
queryDescriptorsQuery set of descriptors.
matchesMatches. Each matches[i] is k or less matches for the same query descriptor.
kCount of best matches found per each query descriptor or less if a query descriptor has less than k possible matches in total.
masksSet of masks. Each masks[i] specifies permissible matches between the input query descriptors and stored train descriptors from the i-th image trainDescCollection[i].
compactResultParameter used when the mask (or masks) is not empty. If compactResult is false, the matches vector has the same size as queryDescriptors rows. If compactResult is true, the matches vector does not contain matches for fully masked-out query descriptors.

◆ knnMatchImpl()

virtual void cv::DescriptorMatcher::knnMatchImpl ( InputArray  queryDescriptors,
std::vector< std::vector< DMatch > > &  matches,
int  k,
InputArrayOfArrays  masks = noArray(),
bool  compactResult = false 
)
protectedpure virtual

In fact the matching is implemented only by the following two methods. These methods suppose that the class object has been trained already. Public match methods call these methods after calling train().

Implemented in cv::BFMatcher, and cv::FlannBasedMatcher.

Here is the call graph for this function:

◆ match() [1/2]

void cv::DescriptorMatcher::match ( InputArray  queryDescriptors,
InputArray  trainDescriptors,
std::vector< DMatch > &  matches,
InputArray  mask = noArray() 
) const
Python:
cv.DescriptorMatcher.match(queryDescriptors, trainDescriptors[, mask]) -> matches
cv.DescriptorMatcher.match(queryDescriptors[, masks]) -> matches

Finds the best match for each descriptor from a query set.

Parameters
queryDescriptorsQuery set of descriptors.
trainDescriptorsTrain set of descriptors. This set is not added to the train descriptors collection stored in the class object.
matchesMatches. If a query descriptor is masked out in mask , no match is added for this descriptor. So, matches size may be smaller than the query descriptors count.
maskMask specifying permissible matches between an input query and train matrices of descriptors.

In the first variant of this method, the train descriptors are passed as an input argument. In the second variant of the method, train descriptors collection that was set by DescriptorMatcher::add is used. Optional mask (or masks) can be passed to specify which query and training descriptors can be matched. Namely, queryDescriptors[i] can be matched with trainDescriptors[j] only if mask.at<uchar>(i,j) is non-zero.

◆ match() [2/2]

void cv::DescriptorMatcher::match ( InputArray  queryDescriptors,
std::vector< DMatch > &  matches,
InputArrayOfArrays  masks = noArray() 
)
Python:
cv.DescriptorMatcher.match(queryDescriptors, trainDescriptors[, mask]) -> matches
cv.DescriptorMatcher.match(queryDescriptors[, masks]) -> matches

This is an overloaded member function, provided for convenience. It differs from the above function only in what argument(s) it accepts.

Parameters
queryDescriptorsQuery set of descriptors.
matchesMatches. If a query descriptor is masked out in mask , no match is added for this descriptor. So, matches size may be smaller than the query descriptors count.
masksSet of masks. Each masks[i] specifies permissible matches between the input query descriptors and stored train descriptors from the i-th image trainDescCollection[i].

◆ radiusMatch() [1/2]

void cv::DescriptorMatcher::radiusMatch ( InputArray  queryDescriptors,
InputArray  trainDescriptors,
std::vector< std::vector< DMatch > > &  matches,
float  maxDistance,
InputArray  mask = noArray(),
bool  compactResult = false 
) const
Python:
cv.DescriptorMatcher.radiusMatch(queryDescriptors, trainDescriptors, maxDistance[, mask[, compactResult]]) -> matches
cv.DescriptorMatcher.radiusMatch(queryDescriptors, maxDistance[, masks[, compactResult]]) -> matches

For each query descriptor, finds the training descriptors not farther than the specified distance.

Parameters
queryDescriptorsQuery set of descriptors.
trainDescriptorsTrain set of descriptors. This set is not added to the train descriptors collection stored in the class object.
matchesFound matches.
compactResultParameter used when the mask (or masks) is not empty. If compactResult is false, the matches vector has the same size as queryDescriptors rows. If compactResult is true, the matches vector does not contain matches for fully masked-out query descriptors.
maxDistanceThreshold for the distance between matched descriptors. Distance means here metric distance (e.g. Hamming distance), not the distance between coordinates (which is measured in Pixels)!
maskMask specifying permissible matches between an input query and train matrices of descriptors.

For each query descriptor, the methods find such training descriptors that the distance between the query descriptor and the training descriptor is equal or smaller than maxDistance. Found matches are returned in the distance increasing order.

◆ radiusMatch() [2/2]

void cv::DescriptorMatcher::radiusMatch ( InputArray  queryDescriptors,
std::vector< std::vector< DMatch > > &  matches,
float  maxDistance,
InputArrayOfArrays  masks = noArray(),
bool  compactResult = false 
)
Python:
cv.DescriptorMatcher.radiusMatch(queryDescriptors, trainDescriptors, maxDistance[, mask[, compactResult]]) -> matches
cv.DescriptorMatcher.radiusMatch(queryDescriptors, maxDistance[, masks[, compactResult]]) -> matches

This is an overloaded member function, provided for convenience. It differs from the above function only in what argument(s) it accepts.

Parameters
queryDescriptorsQuery set of descriptors.
matchesFound matches.
maxDistanceThreshold for the distance between matched descriptors. Distance means here metric distance (e.g. Hamming distance), not the distance between coordinates (which is measured in Pixels)!
masksSet of masks. Each masks[i] specifies permissible matches between the input query descriptors and stored train descriptors from the i-th image trainDescCollection[i].
compactResultParameter used when the mask (or masks) is not empty. If compactResult is false, the matches vector has the same size as queryDescriptors rows. If compactResult is true, the matches vector does not contain matches for fully masked-out query descriptors.

◆ radiusMatchImpl()

virtual void cv::DescriptorMatcher::radiusMatchImpl ( InputArray  queryDescriptors,
std::vector< std::vector< DMatch > > &  matches,
float  maxDistance,
InputArrayOfArrays  masks = noArray(),
bool  compactResult = false 
)
protectedpure virtual

Implemented in cv::BFMatcher, and cv::FlannBasedMatcher.

Here is the call graph for this function:

◆ read() [1/2]

virtual void cv::DescriptorMatcher::read ( const FileNode fn)
virtual
Python:
cv.DescriptorMatcher.read(fileName) -> None
cv.DescriptorMatcher.read(arg1) -> None

Reads algorithm parameters from a file storage.

Reimplemented from cv::Algorithm.

Reimplemented in cv::FlannBasedMatcher.

◆ read() [2/2]

void cv::DescriptorMatcher::read ( const String fileName)
inline
Python:
cv.DescriptorMatcher.read(fileName) -> None
cv.DescriptorMatcher.read(arg1) -> None
Here is the call graph for this function:

◆ train()

virtual void cv::DescriptorMatcher::train ( )
virtual
Python:
cv.DescriptorMatcher.train() -> None

Trains a descriptor matcher.

Trains a descriptor matcher (for example, the flann index). In all methods to match, the method train() is run every time before matching. Some descriptor matchers (for example, BruteForceMatcher) have an empty implementation of this method. Other matchers really train their inner structures (for example, FlannBasedMatcher trains flann::Index ).

Reimplemented in cv::FlannBasedMatcher.

◆ write() [1/4]

void cv::DescriptorMatcher::write ( const Ptr< FileStorage > &  fs,
const String name 
) const
inline
Python:
cv.DescriptorMatcher.write(fileName) -> None
cv.DescriptorMatcher.write(fs, name) -> None

◆ write() [2/4]

void cv::DescriptorMatcher::write ( const String fileName) const
inline
Python:
cv.DescriptorMatcher.write(fileName) -> None
cv.DescriptorMatcher.write(fs, name) -> None
Here is the call graph for this function:

◆ write() [3/4]

virtual void cv::DescriptorMatcher::write ( FileStorage fs) const
virtual
Python:
cv.DescriptorMatcher.write(fileName) -> None
cv.DescriptorMatcher.write(fs, name) -> None

Stores algorithm parameters in a file storage.

Reimplemented from cv::Algorithm.

Reimplemented in cv::FlannBasedMatcher.

◆ write() [4/4]

void cv::DescriptorMatcher::write ( FileStorage fs,
const String name 
) const
inline
Python:
cv.DescriptorMatcher.write(fileName) -> None
cv.DescriptorMatcher.write(fs, name) -> None

Member Data Documentation

◆ trainDescCollection

std::vector<Mat> cv::DescriptorMatcher::trainDescCollection
protected

Collection of descriptors from train images.

◆ utrainDescCollection

std::vector<UMat> cv::DescriptorMatcher::utrainDescCollection
protected

The documentation for this class was generated from the following file: