1 Common Interfaces of Descriptor Matchers
2 ========================================
6 Matchers of keypoint descriptors in OpenCV have wrappers with a common interface that enables you to easily switch
7 between different algorithms solving the same problem. This section is devoted to matching descriptors
8 that are represented as vectors in a multidimensional space. All objects that implement ``vector``
9 descriptor matchers inherit the
10 :ocv:class:`DescriptorMatcher` interface.
14 * An example explaining keypoint matching can be found at opencv_source_code/samples/cpp/descriptor_extractor_matcher.cpp
15 * An example on descriptor matching evaluation can be found at opencv_source_code/samples/cpp/detector_descriptor_matcher_evaluation.cpp
16 * An example on one to many image matching can be found at opencv_source_code/samples/cpp/matching_to_many_images.cpp
20 .. ocv:struct:: DMatch
22 Class for matching keypoint descriptors: query descriptor index,
23 train descriptor index, train image index, and distance between descriptors. ::
27 DMatch() : queryIdx(-1), trainIdx(-1), imgIdx(-1),
28 distance(std::numeric_limits<float>::max()) {}
29 DMatch( int _queryIdx, int _trainIdx, float _distance ) :
30 queryIdx(_queryIdx), trainIdx(_trainIdx), imgIdx(-1),
31 distance(_distance) {}
32 DMatch( int _queryIdx, int _trainIdx, int _imgIdx, float _distance ) :
33 queryIdx(_queryIdx), trainIdx(_trainIdx), imgIdx(_imgIdx),
34 distance(_distance) {}
36 int queryIdx; // query descriptor index
37 int trainIdx; // train descriptor index
38 int imgIdx; // train image index
43 bool operator<( const DMatch &m ) const;
49 .. ocv:class:: DescriptorMatcher : public Algorithm
51 Abstract base class for matching keypoint descriptors. It has two groups
52 of match methods: for matching descriptors of an image with another image or
55 class DescriptorMatcher
58 virtual ~DescriptorMatcher();
60 virtual void add( const vector<Mat>& descriptors );
62 const vector<Mat>& getTrainDescriptors() const;
65 virtual bool isMaskSupported() const = 0;
70 * Group of methods to match descriptors from an image pair.
72 void match( const Mat& queryDescriptors, const Mat& trainDescriptors,
73 vector<DMatch>& matches, const Mat& mask=Mat() ) const;
74 void knnMatch( const Mat& queryDescriptors, const Mat& trainDescriptors,
75 vector<vector<DMatch> >& matches, int k,
76 const Mat& mask=Mat(), bool compactResult=false ) const;
77 void radiusMatch( const Mat& queryDescriptors, const Mat& trainDescriptors,
78 vector<vector<DMatch> >& matches, float maxDistance,
79 const Mat& mask=Mat(), bool compactResult=false ) const;
81 * Group of methods to match descriptors from one image to an image set.
83 void match( const Mat& queryDescriptors, vector<DMatch>& matches,
84 const vector<Mat>& masks=vector<Mat>() );
85 void knnMatch( const Mat& queryDescriptors, vector<vector<DMatch> >& matches,
86 int k, const vector<Mat>& masks=vector<Mat>(),
87 bool compactResult=false );
88 void radiusMatch( const Mat& queryDescriptors, vector<vector<DMatch> >& matches,
89 float maxDistance, const vector<Mat>& masks=vector<Mat>(),
90 bool compactResult=false );
92 virtual void read( const FileNode& );
93 virtual void write( FileStorage& ) const;
95 virtual Ptr<DescriptorMatcher> clone( bool emptyTrainData=false ) const = 0;
97 static Ptr<DescriptorMatcher> create( const string& descriptorMatcherType );
100 vector<Mat> trainDescCollection;
105 DescriptorMatcher::add
106 --------------------------
107 Adds descriptors to train a descriptor collection. If the collection ``trainDescCollectionis`` is not empty, the new descriptors are added to existing train descriptors.
109 .. ocv:function:: void DescriptorMatcher::add( const vector<Mat>& descriptors )
111 :param descriptors: Descriptors to add. Each ``descriptors[i]`` is a set of descriptors from the same train image.
114 DescriptorMatcher::getTrainDescriptors
115 ------------------------------------------
116 Returns a constant link to the train descriptor collection ``trainDescCollection`` .
118 .. ocv:function:: const vector<Mat>& DescriptorMatcher::getTrainDescriptors() const
124 DescriptorMatcher::clear
125 ----------------------------
126 Clears the train descriptor collection.
128 .. ocv:function:: void DescriptorMatcher::clear()
132 DescriptorMatcher::empty
133 ----------------------------
134 Returns true if there are no train descriptors in the collection.
136 .. ocv:function:: bool DescriptorMatcher::empty() const
140 DescriptorMatcher::isMaskSupported
141 --------------------------------------
142 Returns true if the descriptor matcher supports masking permissible matches.
144 .. ocv:function:: bool DescriptorMatcher::isMaskSupported()
148 DescriptorMatcher::train
149 ----------------------------
150 Trains a descriptor matcher
152 .. ocv:function:: void DescriptorMatcher::train()
154 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`` ).
158 DescriptorMatcher::match
159 ----------------------------
160 Finds the best match for each descriptor from a query set.
162 .. ocv:function:: void DescriptorMatcher::match( const Mat& queryDescriptors, const Mat& trainDescriptors, vector<DMatch>& matches, const Mat& mask=Mat() ) const
164 .. ocv:function:: void DescriptorMatcher::match( const Mat& queryDescriptors, vector<DMatch>& matches, const vector<Mat>& masks=vector<Mat>() )
166 :param queryDescriptors: Query set of descriptors.
168 :param trainDescriptors: Train set of descriptors. This set is not added to the train descriptors collection stored in the class object.
170 :param matches: Matches. 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.
172 :param mask: Mask specifying permissible matches between an input query and train matrices of descriptors.
174 :param masks: Set of masks. Each ``masks[i]`` specifies permissible matches between the input query descriptors and stored train descriptors from the i-th image ``trainDescCollection[i]``.
176 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.
180 DescriptorMatcher::knnMatch
181 -------------------------------
182 Finds the k best matches for each descriptor from a query set.
184 .. ocv:function:: void DescriptorMatcher::knnMatch( const Mat& queryDescriptors, const Mat& trainDescriptors, vector<vector<DMatch> >& matches, int k, const Mat& mask=Mat(), bool compactResult=false ) const
186 .. ocv:function:: void DescriptorMatcher::knnMatch( const Mat& queryDescriptors, vector<vector<DMatch> >& matches, int k, const vector<Mat>& masks=vector<Mat>(), bool compactResult=false )
188 :param queryDescriptors: Query set of descriptors.
190 :param trainDescriptors: Train set of descriptors. This set is not added to the train descriptors collection stored in the class object.
192 :param mask: Mask specifying permissible matches between an input query and train matrices of descriptors.
194 :param masks: Set of masks. Each ``masks[i]`` specifies permissible matches between the input query descriptors and stored train descriptors from the i-th image ``trainDescCollection[i]``.
196 :param matches: Matches. Each ``matches[i]`` is k or less matches for the same query descriptor.
198 :param k: Count of best matches found per each query descriptor or less if a query descriptor has less than k possible matches in total.
200 :param compactResult: Parameter 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.
202 These extended variants of :ocv:func:`DescriptorMatcher::match` methods find several best matches for each query descriptor. The matches are returned in the distance increasing order. See :ocv:func:`DescriptorMatcher::match` for the details about query and train descriptors.
206 DescriptorMatcher::radiusMatch
207 ----------------------------------
208 For each query descriptor, finds the training descriptors not farther than the specified distance.
210 .. ocv:function:: void DescriptorMatcher::radiusMatch( const Mat& queryDescriptors, const Mat& trainDescriptors, vector<vector<DMatch> >& matches, float maxDistance, const Mat& mask=Mat(), bool compactResult=false ) const
212 .. ocv:function:: void DescriptorMatcher::radiusMatch( const Mat& queryDescriptors, vector<vector<DMatch> >& matches, float maxDistance, const vector<Mat>& masks=vector<Mat>(), bool compactResult=false )
214 :param queryDescriptors: Query set of descriptors.
216 :param trainDescriptors: Train set of descriptors. This set is not added to the train descriptors collection stored in the class object.
218 :param mask: Mask specifying permissible matches between an input query and train matrices of descriptors.
220 :param masks: Set of masks. Each ``masks[i]`` specifies permissible matches between the input query descriptors and stored train descriptors from the i-th image ``trainDescCollection[i]``.
222 :param matches: Found matches.
224 :param compactResult: Parameter 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.
226 :param maxDistance: Threshold 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)!
228 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.
232 DescriptorMatcher::clone
233 ----------------------------
236 .. ocv:function:: Ptr<DescriptorMatcher> DescriptorMatcher::clone( bool emptyTrainData=false )
238 :param emptyTrainData: If ``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.
242 DescriptorMatcher::create
243 -----------------------------
244 Creates a descriptor matcher of a given type with the default parameters (using default constructor).
246 .. ocv:function:: Ptr<DescriptorMatcher> DescriptorMatcher::create( const string& descriptorMatcherType )
248 :param descriptorMatcherType: Descriptor matcher type. Now the following matcher types are supported:
251 ``BruteForce`` (it uses ``L2`` )
255 ``BruteForce-Hamming``
257 ``BruteForce-Hamming(2)``
267 .. ocv:class:: BFMatcher : public DescriptorMatcher
269 Brute-force descriptor matcher. For each descriptor in the first set, this matcher finds the closest descriptor in the second set by trying each one. This descriptor matcher supports masking permissible matches of descriptor sets.
274 Brute-force matcher constructor.
276 .. ocv:function:: BFMatcher::BFMatcher( int normType=NORM_L2, bool crossCheck=false )
278 :param normType: One of ``NORM_L1``, ``NORM_L2``, ``NORM_HAMMING``, ``NORM_HAMMING2``. ``L1`` and ``L2`` norms are preferable choices for SIFT and SURF descriptors, ``NORM_HAMMING`` should be used with ORB, BRISK and BRIEF, ``NORM_HAMMING2`` should be used with ORB when ``WTA_K==3`` or ``4`` (see ORB::ORB constructor description).
280 :param crossCheck: If it is false, this is will be default BFMatcher behaviour when it finds the k nearest neighbors for each query descriptor. If ``crossCheck==true``, then the ``knnMatch()`` method with ``k=1`` will only return pairs ``(i,j)`` such that for ``i-th`` query descriptor the ``j-th`` descriptor in the matcher's collection is the nearest and vice versa, i.e. the ``BFMathcher`` will only return consistent pairs. Such technique usually produces best results with minimal number of outliers when there are enough matches. This is alternative to the ratio test, used by D. Lowe in SIFT paper.
285 .. ocv:class:: FlannBasedMatcher : public DescriptorMatcher
287 Flann-based descriptor matcher. This matcher trains :ocv:class:`flann::Index_` on a train descriptor collection and calls its nearest search methods to find the best matches. So, this matcher may be faster when matching a large train collection than the brute force matcher. ``FlannBasedMatcher`` does not support masking permissible matches of descriptor sets because ``flann::Index`` does not support this. ::
289 class FlannBasedMatcher : public DescriptorMatcher
293 const Ptr<flann::IndexParams>& indexParams=new flann::KDTreeIndexParams(),
294 const Ptr<flann::SearchParams>& searchParams=new flann::SearchParams() );
296 virtual void add( const vector<Mat>& descriptors );
297 virtual void clear();
299 virtual void train();
300 virtual bool isMaskSupported() const;
302 virtual Ptr<DescriptorMatcher> clone( bool emptyTrainData=false ) const;