Mat img2 = imread(argv[2], IMREAD_GRAYSCALE);
#.
- Detect keypoints in both images. ::
+ Detect keypoints in both images and compute descriptors for each of the keypoints. ::
// detecting keypoints
- FastFeatureDetector detector(15);
+ Ptr<Feature2D> surf = SURF::create();
vector<KeyPoint> keypoints1;
- detector.detect(img1, keypoints1);
-
- ... // do the same for the second image
-
-#.
- Compute descriptors for each of the keypoints. ::
-
- // computing descriptors
- SurfDescriptorExtractor extractor;
Mat descriptors1;
- extractor.compute(img1, keypoints1, descriptors1);
+ surf->detectAndCompute(img1, Mat(), keypoints1, descriptors1);
- ... // process keypoints from the second image as well
+ ... // do the same for the second image
#.
Now, find the closest matches between descriptors from the first image to the second: ::
We load two images and check if they are loaded correctly.::
// detecting keypoints
- FastFeatureDetector detector(15);
+ Ptr<FeatureDetector> detector = FastFeatureDetector::create(15);
vector<KeyPoint> keypoints1, keypoints2;
- detector.detect(img1, keypoints1);
- detector.detect(img2, keypoints2);
+ detector->detect(img1, keypoints1);
+ detector->detect(img2, keypoints2);
First, we create an instance of a keypoint detector. All detectors inherit the abstract ``FeatureDetector`` interface, but the constructors are algorithm-dependent. The first argument to each detector usually controls the balance between the amount of keypoints and their stability. The range of values is different for different detectors (For instance, *FAST* threshold has the meaning of pixel intensity difference and usually varies in the region *[0,40]*. *SURF* threshold is applied to a Hessian of an image and usually takes on values larger than *100*), so use defaults in case of doubt. ::
// computing descriptors
- SurfDescriptorExtractor extractor;
+ Ptr<SURF> extractor = SURF::create();
Mat descriptors1, descriptors2;
- extractor.compute(img1, keypoints1, descriptors1);
- extractor.compute(img2, keypoints2, descriptors2);
+ extractor->compute(img1, keypoints1, descriptors1);
+ extractor->compute(img2, keypoints2, descriptors2);
We create an instance of descriptor extractor. The most of OpenCV descriptors inherit ``DescriptorExtractor`` abstract interface. Then we compute descriptors for each of the keypoints. The output ``Mat`` of the ``DescriptorExtractor::compute`` method contains a descriptor in a row *i* for each *i*-th keypoint. Note that the method can modify the keypoints vector by removing the keypoints such that a descriptor for them is not defined (usually these are the keypoints near image border). The method makes sure that the ouptut keypoints and descriptors are consistent with each other (so that the number of keypoints is equal to the descriptors row count). ::
//! finds circles' grid pattern of the specified size in the image
CV_EXPORTS_W bool findCirclesGrid( InputArray image, Size patternSize,
OutputArray centers, int flags = CALIB_CB_SYMMETRIC_GRID,
- const Ptr<FeatureDetector> &blobDetector = makePtr<SimpleBlobDetector>());
+ const Ptr<FeatureDetector> &blobDetector = SimpleBlobDetector::create());
//! finds intrinsic and extrinsic camera parameters from several fews of a known calibration pattern.
CV_EXPORTS_W double calibrateCamera( InputArrayOfArrays objectPoints,
cv::Mat R, rvec = _rvec.getMat(), tvec = _tvec.getMat();
double f = PnP.compute_pose(R, tvec);
cv::Rodrigues(R, rvec);
- cameraMatrix.at<double>(0,0) = cameraMatrix.at<double>(1,1) = f;
+ if(cameraMatrix.type() == CV_32F)
+ cameraMatrix.at<float>(0,0) = cameraMatrix.at<float>(1,1) = f;
+ else
+ cameraMatrix.at<double>(0,0) = cameraMatrix.at<double>(1,1) = f;
return true;
}
else
:param masks: Masks for each input image specifying where to look for keypoints (optional). ``masks[i]`` is a mask for ``images[i]``.
-FeatureDetector::create
------------------------
-Creates a feature detector by its name.
-
-.. ocv:function:: Ptr<FeatureDetector> FeatureDetector::create( const String& detectorType )
-
-.. ocv:pyfunction:: cv2.FeatureDetector_create(detectorType) -> retval
-
- :param detectorType: Feature detector type.
-
-The following detector types are supported:
-
-* ``"FAST"`` -- :ocv:class:`FastFeatureDetector`
-* ``"ORB"`` -- :ocv:class:`ORB`
-* ``"BRISK"`` -- :ocv:class:`BRISK`
-* ``"MSER"`` -- :ocv:class:`MSER`
-* ``"GFTT"`` -- :ocv:class:`GoodFeaturesToTrackDetector`
-* ``"HARRIS"`` -- :ocv:class:`GoodFeaturesToTrackDetector` with Harris detector enabled
-* ``"SimpleBlob"`` -- :ocv:class:`SimpleBlobDetector`
-
FastFeatureDetector
-------------------
-.. ocv:class:: FastFeatureDetector : public FeatureDetector
+.. ocv:class:: FastFeatureDetector : public Feature2D
Wrapping class for feature detection using the
:ocv:func:`FAST` method. ::
- class FastFeatureDetector : public FeatureDetector
+ class FastFeatureDetector : public Feature2D
{
public:
- FastFeatureDetector( int threshold=1, bool nonmaxSuppression=true, type=FastFeatureDetector::TYPE_9_16 );
- virtual void read( const FileNode& fn );
- virtual void write( FileStorage& fs ) const;
- protected:
- ...
+ static Ptr<FastFeatureDetector> create( int threshold=1, bool nonmaxSuppression=true, type=FastFeatureDetector::TYPE_9_16 );
};
-GoodFeaturesToTrackDetector
+GFTTDetector
---------------------------
-.. ocv:class:: GoodFeaturesToTrackDetector : public FeatureDetector
+.. ocv:class:: GFTTDetector : public FeatureDetector
Wrapping class for feature detection using the
:ocv:func:`goodFeaturesToTrack` function. ::
- class GoodFeaturesToTrackDetector : public FeatureDetector
+ class GFTTDetector : public Feature2D
{
public:
- class Params
- {
- public:
- Params( int maxCorners=1000, double qualityLevel=0.01,
- double minDistance=1., int blockSize=3,
- bool useHarrisDetector=false, double k=0.04 );
- void read( const FileNode& fn );
- void write( FileStorage& fs ) const;
-
- int maxCorners;
- double qualityLevel;
- double minDistance;
- int blockSize;
- bool useHarrisDetector;
- double k;
- };
-
- GoodFeaturesToTrackDetector( const GoodFeaturesToTrackDetector::Params& params=
- GoodFeaturesToTrackDetector::Params() );
- GoodFeaturesToTrackDetector( int maxCorners, double qualityLevel,
- double minDistance, int blockSize=3,
- bool useHarrisDetector=false, double k=0.04 );
- virtual void read( const FileNode& fn );
- virtual void write( FileStorage& fs ) const;
- protected:
- ...
+ enum { USE_HARRIS_DETECTOR=10000 };
+ static Ptr<GFTTDetector> create( int maxCorners=1000, double qualityLevel=0.01,
+ double minDistance=1, int blockSize=3,
+ bool useHarrisDetector=false, double k=0.04 );
};
-MserFeatureDetector
+MSER
-------------------
-.. ocv:class:: MserFeatureDetector : public FeatureDetector
+.. ocv:class:: MSER : public Feature2D
-Wrapping class for feature detection using the
-:ocv:class:`MSER` class. ::
+Maximally stable region detector ::
- class MserFeatureDetector : public FeatureDetector
+ class MSER : public Feature2D
{
public:
- MserFeatureDetector( CvMSERParams params=cvMSERParams() );
- MserFeatureDetector( int delta, int minArea, int maxArea,
- double maxVariation, double minDiversity,
- int maxEvolution, double areaThreshold,
- double minMargin, int edgeBlurSize );
- virtual void read( const FileNode& fn );
- virtual void write( FileStorage& fs ) const;
- protected:
- ...
+ enum
+ {
+ DELTA=10000, MIN_AREA=10001, MAX_AREA=10002, PASS2_ONLY=10003,
+ MAX_EVOLUTION=10004, AREA_THRESHOLD=10005,
+ MIN_MARGIN=10006, EDGE_BLUR_SIZE=10007
+ };
+
+ //! the full constructor
+ static Ptr<MSER> create( int _delta=5, int _min_area=60, int _max_area=14400,
+ double _max_variation=0.25, double _min_diversity=.2,
+ int _max_evolution=200, double _area_threshold=1.01,
+ double _min_margin=0.003, int _edge_blur_size=5 );
+
+ virtual void detectRegions( InputArray image,
+ std::vector<std::vector<Point> >& msers,
+ std::vector<Rect>& bboxes ) = 0;
};
SimpleBlobDetector
float minConvexity, maxConvexity;
};
- SimpleBlobDetector(const SimpleBlobDetector::Params ¶meters = SimpleBlobDetector::Params());
-
- protected:
- ...
+ static Ptr<SimpleBlobDetector> create(const SimpleBlobDetector::Params
+ ¶meters = SimpleBlobDetector::Params());
};
The class implements a simple algorithm for extracting blobs from an image:
.. ocv:function:: void FAST( InputArray image, vector<KeyPoint>& keypoints, int threshold, bool nonmaxSuppression=true )
.. ocv:function:: void FAST( InputArray image, vector<KeyPoint>& keypoints, int threshold, bool nonmaxSuppression, int type )
-.. ocv:pyfunction:: cv2.FastFeatureDetector([, threshold[, nonmaxSuppression]]) -> <FastFeatureDetector object>
-.. ocv:pyfunction:: cv2.FastFeatureDetector(threshold, nonmaxSuppression, type) -> <FastFeatureDetector object>
-.. ocv:pyfunction:: cv2.FastFeatureDetector.detect(image[, mask]) -> keypoints
-
-
:param image: grayscale image where keypoints (corners) are detected.
:param keypoints: keypoints detected on the image.
// runs the extractor on the specified image; returns the MSERs,
// each encoded as a contour (vector<Point>, see findContours)
// the optional mask marks the area where MSERs are searched for
- void operator()( const Mat& image, vector<vector<Point> >& msers, const Mat& mask ) const;
+ void detectRegions( InputArray image, vector<vector<Point> >& msers, vector<Rect>& bboxes ) const;
};
The class encapsulates all the parameters of the MSER extraction algorithm (see
{
public:
// the size of the signature in bytes
- enum { kBytes = 32, HARRIS_SCORE=0, FAST_SCORE=1 };
+ enum
+ {
+ kBytes = 32, HARRIS_SCORE=0, FAST_SCORE=1,
+ NFEATURES=10000, SCALE_FACTOR=10001, NLEVELS=10002,
+ EDGE_THRESHOLD=10003, FIRST_LEVEL=10004, WTA_K=10005,
+ SCORE_TYPE=10006, PATCH_SIZE=10007, FAST_THRESHOLD=10008
+ };
CV_WRAP static Ptr<ORB> create(int nfeatures = 500, float scaleFactor = 1.2f, int nlevels = 8, int edgeThreshold = 31,
int firstLevel = 0, int WTA_K=2, int scoreType=ORB::HARRIS_SCORE, int patchSize=31, int fastThreshold = 20);
{
}
+ void set(int prop, double value)
+ {
+ if( prop == USE_HARRIS_DETECTOR )
+ useHarrisDetector = value != 0;
+ else
+ CV_Error(Error::StsBadArg, "");
+ }
+
+ double get(int prop) const
+ {
+ double value = 0;
+ if( prop == USE_HARRIS_DETECTOR )
+ value = useHarrisDetector;
+ else
+ CV_Error(Error::StsBadArg, "");
+ return value;
+ }
+
void detect( InputArray _image, std::vector<KeyPoint>& keypoints, InputArray _mask )
{
std::vector<Point2f> corners;
static bool
ocl_computeOrbDescriptors(const UMat& imgbuf, const UMat& layerInfo,
const UMat& keypoints, UMat& desc, const UMat& pattern,
- int nkeypoints, int dsize, int WTA_K)
+ int nkeypoints, int dsize, int wta_k)
{
size_t globalSize[] = {nkeypoints};
ocl::Kernel desc_ker("ORB_computeDescriptor", ocl::features2d::orb_oclsrc,
- format("-D ORB_DESCRIPTORS -D WTA_K=%d", WTA_K));
+ format("-D ORB_DESCRIPTORS -D wta_k=%d", wta_k));
if( desc_ker.empty() )
return false;
static void
computeOrbDescriptors( const Mat& imagePyramid, const std::vector<Rect>& layerInfo,
const std::vector<float>& layerScale, std::vector<KeyPoint>& keypoints,
- Mat& descriptors, const std::vector<Point>& _pattern, int dsize, int WTA_K )
+ Mat& descriptors, const std::vector<Point>& _pattern, int dsize, int wta_k )
{
int step = (int)imagePyramid.step;
int j, i, nkeypoints = (int)keypoints.size();
center[iy*step + ix+1]*x*(1-y) + center[(iy+1)*step + ix+1]*x*y))
#endif
- if( WTA_K == 2 )
+ if( wta_k == 2 )
{
for (i = 0; i < dsize; ++i, pattern += 16)
{
desc[i] = (uchar)val;
}
}
- else if( WTA_K == 3 )
+ else if( wta_k == 3 )
{
for (i = 0; i < dsize; ++i, pattern += 12)
{
desc[i] = (uchar)val;
}
}
- else if( WTA_K == 4 )
+ else if( wta_k == 4 )
{
for (i = 0; i < dsize; ++i, pattern += 16)
{
}
}
else
- CV_Error( Error::StsBadSize, "Wrong WTA_K. It can be only 2, 3 or 4." );
+ CV_Error( Error::StsBadSize, "Wrong wta_k. It can be only 2, 3 or 4." );
#undef GET_VALUE
}
}
explicit ORB_Impl(int _nfeatures, float _scaleFactor, int _nlevels, int _edgeThreshold,
int _firstLevel, int _WTA_K, int _scoreType, int _patchSize, int _fastThreshold) :
nfeatures(_nfeatures), scaleFactor(_scaleFactor), nlevels(_nlevels),
- edgeThreshold(_edgeThreshold), firstLevel(_firstLevel), WTA_K(_WTA_K),
+ edgeThreshold(_edgeThreshold), firstLevel(_firstLevel), wta_k(_WTA_K),
scoreType(_scoreType), patchSize(_patchSize), fastThreshold(_fastThreshold)
{}
+ void set(int prop, double value)
+ {
+ if( prop == NFEATURES )
+ nfeatures = cvRound(value);
+ else if( prop == SCALE_FACTOR )
+ scaleFactor = value;
+ else if( prop == NLEVELS )
+ nlevels = cvRound(value);
+ else if( prop == EDGE_THRESHOLD )
+ edgeThreshold = cvRound(value);
+ else if( prop == FIRST_LEVEL )
+ firstLevel = cvRound(value);
+ else if( prop == WTA_K )
+ wta_k = cvRound(value);
+ else if( prop == SCORE_TYPE )
+ scoreType = cvRound(value);
+ else if( prop == PATCH_SIZE )
+ patchSize = cvRound(value);
+ else if( prop == FAST_THRESHOLD )
+ fastThreshold = cvRound(value);
+ else
+ CV_Error(Error::StsBadArg, "");
+ }
+
+ double get(int prop) const
+ {
+ double value = 0;
+ if( prop == NFEATURES )
+ value = nfeatures;
+ else if( prop == SCALE_FACTOR )
+ value = scaleFactor;
+ else if( prop == NLEVELS )
+ value = nlevels;
+ else if( prop == EDGE_THRESHOLD )
+ value = edgeThreshold;
+ else if( prop == FIRST_LEVEL )
+ value = firstLevel;
+ else if( prop == WTA_K )
+ value = wta_k;
+ else if( prop == SCORE_TYPE )
+ value = scoreType;
+ else if( prop == PATCH_SIZE )
+ value = patchSize;
+ else if( prop == FAST_THRESHOLD )
+ value = fastThreshold;
+ else
+ CV_Error(Error::StsBadArg, "");
+ return value;
+ }
+
// returns the descriptor size in bytes
int descriptorSize() const;
// returns the descriptor type
int nlevels;
int edgeThreshold;
int firstLevel;
- int WTA_K;
+ int wta_k;
int scoreType;
int patchSize;
int fastThreshold;
makeRandomPattern(patchSize, patternbuf, npoints);
}
- CV_Assert( WTA_K == 2 || WTA_K == 3 || WTA_K == 4 );
+ CV_Assert( wta_k == 2 || wta_k == 3 || wta_k == 4 );
- if( WTA_K == 2 )
+ if( wta_k == 2 )
std::copy(pattern0, pattern0 + npoints, std::back_inserter(pattern));
else
{
int ntuples = descriptorSize()*4;
- initializeOrbPattern(pattern0, pattern, ntuples, WTA_K, npoints);
+ initializeOrbPattern(pattern0, pattern, ntuples, wta_k, npoints);
}
for( level = 0; level < nLevels; level++ )
UMat udescriptors = _descriptors.getUMat();
useOCL = ocl_computeOrbDescriptors(uimagePyramid, ulayerInfo,
ukeypoints, udescriptors, upattern,
- nkeypoints, dsize, WTA_K);
+ nkeypoints, dsize, wta_k);
if(useOCL)
{
CV_IMPL_ADD(CV_IMPL_OCL);
{
Mat descriptors = _descriptors.getMat();
computeOrbDescriptors(imagePyramid, layerInfo, layerScale,
- keypoints, descriptors, pattern, dsize, WTA_K);
+ keypoints, descriptors, pattern, dsize, wta_k);
}
}
}
Ptr<ORB> ORB::create(int nfeatures, float scaleFactor, int nlevels, int edgeThreshold,
- int firstLevel, int WTA_K, int scoreType, int patchSize, int fastThreshold)
+ int firstLevel, int wta_k, int scoreType, int patchSize, int fastThreshold)
{
return makePtr<ORB_Impl>(nfeatures, scaleFactor, nlevels, edgeThreshold,
- firstLevel, WTA_K, scoreType, patchSize, fastThreshold);
+ firstLevel, wta_k, scoreType, patchSize, fastThreshold);
}
}
TEST( Features2d_Detector_Harris, regression )
{
- CV_FeatureDetectorTest test( "detector-harris", GFTTDetector::create(1000, 0.01, 1, 3, true, 0.04));
+ Ptr<FeatureDetector> gftt = GFTTDetector::create();
+ gftt->set(GFTTDetector::USE_HARRIS_DETECTOR, 1);
+ CV_FeatureDetectorTest test( "detector-harris", gftt);
test.safe_run();
}
TEST(Features2d_Detector_Keypoints_HARRIS, validation)
{
+
CV_FeatureDetectorKeypointsTest test(GFTTDetector::create(1000, 0.01, 1, 3, true, 0.04));
test.safe_run();
}
TEST(Features2d_Detector_Keypoints_GFTT, validation)
{
- CV_FeatureDetectorKeypointsTest test(GFTTDetector::create());
+ Ptr<FeatureDetector> gftt = GFTTDetector::create();
+ gftt->set(GFTTDetector::USE_HARRIS_DETECTOR, 1);
+ CV_FeatureDetectorKeypointsTest test(gftt);
test.safe_run();
}
break;
case HARRIS:
fd = GFTTDetector::create();
+ fd->set(GFTTDetector::USE_HARRIS_DETECTOR, 1);
break;
case SIMPLEBLOB:
fd = SimpleBlobDetector::create();
self.assertEqual(cv2.countNonZero(inliers), pattern_size[0]*pattern_size[1])
def test_fast(self):
- fd = cv2.FastFeatureDetector(30, True)
+ fd = cv2.FastFeatureDetector_create(30, True)
img = self.get_sample("samples/cpp/right02.jpg", 0)
img = cv2.medianBlur(img, 3)
imgc = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)
#include "opencv2/xfeatures2d.hpp"
#endif
+using xfeatures2d::SURF;
+
namespace {
struct DistIdxPair
{
if (num_octaves_descr == num_octaves && num_layers_descr == num_layers)
{
- surf = xfeatures2d::SURF::create();
+ surf = SURF::create();
if( !surf )
CV_Error( Error::StsNotImplemented, "OpenCV was built without SURF support" );
- surf->set("hessianThreshold", hess_thresh);
- surf->set("nOctaves", num_octaves);
- surf->set("nOctaveLayers", num_layers);
+ surf->set(SURF::HESSIAN_THRESHOLD, hess_thresh);
+ surf->set(SURF::NOCTAVES, num_octaves);
+ surf->set(SURF::NOCTAVE_LAYERS, num_layers);
}
else
{
- detector_ = xfeatures2d::SURF::create();
- extractor_ = xfeatures2d::SURF::create();
+ detector_ = SURF::create();
+ extractor_ = SURF::create();
if( !detector_ || !extractor_ )
CV_Error( Error::StsNotImplemented, "OpenCV was built without SURF support" );
- detector_->set("hessianThreshold", hess_thresh);
- detector_->set("nOctaves", num_octaves);
- detector_->set("nOctaveLayers", num_layers);
+ detector_->set(SURF::HESSIAN_THRESHOLD, hess_thresh);
+ detector_->set(SURF::NOCTAVES, num_octaves);
+ detector_->set(SURF::NOCTAVE_LAYERS, num_layers);
- extractor_->set("nOctaves", num_octaves_descr);
- extractor_->set("nOctaveLayers", num_layers_descr);
+ extractor_->set(SURF::NOCTAVES, num_octaves_descr);
+ extractor_->set(SURF::NOCTAVE_LAYERS, num_layers_descr);
}
}
Mat& mRgb = *(Mat*)addrRgba;
vector<KeyPoint> v;
- FastFeatureDetector detector(50);
- detector.detect(mGr, v);
+ Ptr<FeatureDetector> detector = FastFeatureDetector::create(50);
+ detector->detect(mGr, v);
for( unsigned int i = 0; i < v.size(); i++ )
{
const KeyPoint& kp = v[i];
{
cv::Mat result;
cv::Mat intermediateMat;
- cv::FastFeatureDetector detector(50);
+ cv::Ptr<cv::FeatureDetector> detector = cv::FastFeatureDetector::create(50);
std::vector<cv::KeyPoint> features;
image.copyTo(result);
cv::cvtColor(image, intermediateMat, CV_RGBA2GRAY);
- detector.detect(intermediateMat, features);
+ detector->detect(intermediateMat, features);
for( unsigned int i = 0; i < std::min(features.size(), (size_t)50); i++ )
{