KeyPoint¶

class KeyPoint(*args, **kwargs)¶

Bases: BasicKeyPoint

Class that allows keypoints detection (and descriptors extraction) and matching thanks to OpenCV library. Thus to enable this class OpenCV should be installed. Installation instructions are provided here https://visp.inria.fr/3rd_opencv .

This class permits to use different types of detectors, extractors and matchers easily. So, the classical SIFT and SURF keypoints could be used, as well as ORB, FAST, (etc.) keypoints, depending of the version of OpenCV you use.

Note

Due to some patents, SIFT and SURF are packaged in an external module called nonfree module in OpenCV version before 3.0.0 and in xfeatures2d from 3.0.0. You have to check you have the corresponding module to use SIFT and SURF.

The goal of this class is to provide a tool to match reference keypoints from a reference image (or train keypoints in OpenCV terminology) and detected keypoints from a current image (or query keypoints in OpenCV terminology).

If you supply the corresponding 3D coordinates corresponding to the 2D coordinates of the reference keypoints, you can also estimate the pose of the object by matching a set of detected keypoints in the current image with the reference keypoints.

If you use this class, the first thing you have to do is to build the reference keypoints by detecting keypoints in a reference image which contains the object to detect. Then you match keypoints detected in a current image with those detected in a reference image by calling matchPoint() methods. You can access to the lists of matched points thanks to the methods getMatchedPointsInReferenceImage() and getMatchedPointsInCurrentImage(). These two methods return a list of matched points. The nth element of the first list is matched with the nth element of the second list. To provide easy compatibility with OpenCV terminology, getTrainKeyPoints() give you access to the list of keypoints detected in train images (or reference images) and getQueryKeyPoints() give you access to the list of keypoints detected in a query image (or current image). The method getMatches() give you access to a list of cv::DMatch with the correspondence between the index of the train keypoints and the index of the query keypoints.

The following small example shows how to use the class to do the matching between current and reference keypoints.

#include <visp3/core/vpImage.h>
#include <visp3/vision/vpKeyPoint.h>

#ifdef ENABLE_VISP_NAMESPACE
using namespace VISP_NAMESPACE_NAME;
#endif

int main()
{
#if (VISP_HAVE_OPENCV_VERSION >= 0x020300)
  vpImage<unsigned char> Irefrence;
  vpImage<unsigned char> Icurrent;

  vpKeyPoint::vpFilterMatchingType filterType = vpKeyPoint::ratioDistanceThreshold;
  vpKeyPoint keypoint("ORB", "ORB", "BruteForce-Hamming", filterType);

  // First grab the reference image Irefrence
  // Add your code to load the reference image in Ireference

  // Build the reference ORB points.
  keypoint.buildReference(Irefrence);

  // Then grab another image which represents the current image Icurrent

  // Match points between the reference points and the ORB points computed in the current image.
  keypoint.matchPoint(Icurrent);

  // Display the matched points
  keypoint.display(Irefrence, Icurrent);
#endif

  return (0);
}

It is also possible to build the reference keypoints in a region of interest (ROI) of an image and find keypoints to match in only a part of the current image. The small following example shows how to do this:

#include <visp3/core/vpDisplay.h>
#include <visp3/core/vpImage.h>
#include <visp3/vision/vpKeyPoint.h>

#ifdef ENABLE_VISP_NAMESPACE
using namespace VISP_NAMESPACE_NAME;
#endif

int main()
{
#if (VISP_HAVE_OPENCV_VERSION >= 0x020300)
  vpImage<unsigned char> Ireference;
  vpImage<unsigned char> Icurrent;

  vpKeyPoint::vpFilterMatchingType filterType = vpKeyPoint::ratioDistanceThreshold;
  vpKeyPoint keypoint("ORB", "ORB", "BruteForce-Hamming", filterType);

  //First grab the reference image Irefrence
  //Add your code to load the reference image in Ireference

  //Select a part of the image by clicking on two points which define a rectangle
  vpImagePoint corners[2];
  for (int i=0 ; i < 2 ; i++) {
    vpDisplay::getClick(Ireference, corners[i]);
  }

  //Build the reference ORB points.
  int nbrRef;
  unsigned int height, width;
  height = (unsigned int)(corners[1].get_i() - corners[0].get_i());
  width = (unsigned int)(corners[1].get_j() - corners[0].get_j());
  nbrRef = keypoint.buildReference(Ireference, corners[0], height, width);

  //Then grab another image which represents the current image Icurrent

  //Select a part of the image by clicking on two points which define a rectangle
  for (int i=0 ; i < 2 ; i++) {
    vpDisplay::getClick(Icurrent, corners[i]);
  }

  //Match points between the reference points and the ORB points computed in the current image.
  int nbrMatched;
  height = (unsigned int)(corners[1].get_i() - corners[0].get_i());
  width = (unsigned int)(corners[1].get_j() - corners[0].get_j());
  nbrMatched = keypoint.matchPoint(Icurrent, corners[0], height, width);

  //Display the matched points
  keypoint.display(Ireference, Icurrent);
#endif

  return(0);
}

This class is also described in tutorial-matching.

Overloaded function.

__init__(self: visp._visp.vision.KeyPoint, detectorType: visp._visp.vision.KeyPoint.FeatureDetectorType, descriptorType: visp._visp.vision.KeyPoint.FeatureDescriptorType, matcherName: str, filterType: visp._visp.vision.KeyPoint.FilterMatchingType = ratioDistanceThreshold) -> None
__init__(self: visp._visp.vision.KeyPoint, detectorName: str = ORB, extractorName: str = ORB, matcherName: str = BruteForce-Hamming, filterType: visp._visp.vision.KeyPoint.FilterMatchingType = ratioDistanceThreshold) -> None
__init__(self: visp._visp.vision.KeyPoint, detectorNames: list[str], extractorNames: list[str], matcherName: str = BruteForce, filterType: visp._visp.vision.KeyPoint.FilterMatchingType = ratioDistanceThreshold) -> None

Methods

`__init__`	Overloaded function.
`buildReference`	Overloaded function.
`compute3D`	Overloaded function.
`createImageMatching`	Overloaded function.
`detect`	Overloaded function.
`display`	Overloaded function.
`displayMatching`	Overloaded function.
`getCovarianceMatrix`	Get the covariance matrix when estimating the pose using the Virtual Visual Servoing approach.
`getDetectionTime`	Get the elapsed time to compute the keypoint detection.
`getDetector`	Overloaded function.
`getDetectorNames`	Get the feature detector name associated to the type.
`getExtractionTime`	Get the elapsed time to compute the keypoint extraction.
`getExtractor`	Overloaded function.
`getExtractorNames`	Get the feature descriptor extractor name associated to the type.
`getImageFormat`	Get the image format to use when saving training images.
`getMatchQueryToTrainKeyPoints`	Get the list of pairs with the correspondence between the matched query and train keypoints.
`getMatcher`	Get the matcher pointer.
`getMatches`	Get the list of matches (correspondences between the indexes of the detected keypoints and the train keypoints).
`getMatchingTime`	Get the elapsed time to compute the matching.
`getNbImages`	Get the number of train images.
`getObjectPoints`	Overloaded function.
`getPoseTime`	Get the elapsed time to compute the pose.
`getQueryDescriptors`	Get the descriptors matrix for the query keypoints.
`getQueryKeyPoints`	Overloaded function.
`getRansacInliers`	Get the list of Ransac inliers.
`getRansacOutliers`	Get the list of Ransac outliers.
`getTrainDescriptors`	Get the train descriptors matrix.
`getTrainKeyPoints`	Overloaded function.
`getTrainPoints`	Overloaded function.
`initMatcher`	Initialize a matcher based on its name.
`insertImageMatching`	Overloaded function.
`loadConfigFile`	Load configuration parameters from an XML config file.
`loadLearningData`	Load learning data saved on disk.
`match`	Match keypoints based on distance between their descriptors.
`matchPoint`	Overloaded function.
`reset`	Reset the instance as if we would declare another vpKeyPoint variable.
`saveLearningData`	Save the learning data in a file in XML or binary mode.
`setCovarianceComputation`	Set if the covariance matrix has to be computed in the Virtual Visual Servoing approach.
`setDetectionMethod`
`setDetector`	Overloaded function.
`setDetectors`	Set and initialize a list of detectors denominated by their names detectorNames .
`setExtractor`	Overloaded function.
`setExtractors`	Set and initialize a list of extractors denominated by their names extractorNames .
`setFilterMatchingType`
`setImageFormat`
`setMatcher`	Set and initialize a matcher denominated by his name matcherName .
`setMatchingFactorThreshold`	Set the factor value for the filtering method: constantFactorDistanceThreshold.
`setMatchingRatioThreshold`	Set the ratio value for the filtering method: ratioDistanceThreshold.
`setMaxFeatures`	Set maximum number of keypoints to extract.
`setRansacConsensusPercentage`	Set the percentage value for defining the cardinality of the consensus group.
`setRansacFilterFlag`	Set filter flag for RANSAC pose estimation.
`setRansacIteration`	Set the maximum number of iterations for the Ransac pose estimation method.
`setRansacMinInlierCount`	Set the minimum number of inlier for the Ransac pose estimation method.
`setRansacParallel`	Use or not the multithreaded version.
`setRansacParallelNbThreads`	Set the number of threads to use if multithreaded RANSAC pose.
`setRansacReprojectionError`	Set the maximum reprojection error (in pixel) to determine if a point is an inlier or not.
`setRansacThreshold`	Set the maximum error (in meter) to determine if a point is an inlier or not.
`setUseAffineDetection`	Set if multiple affine transformations must be used to detect and extract keypoints.
`setUseMatchTrainToQuery`	Set if we want to match the train keypoints to the query keypoints.
`setUseRansacConsensusPercentage`	Set the flag to choose between a percentage value of inliers for the cardinality of the consensus group or a minimum number.
`setUseRansacVVS`	Set the flag to choose between the OpenCV or ViSP Ransac pose estimation function.
`setUseSingleMatchFilter`	Set the flag to filter matches where multiple query keypoints are matched to the same train keypoints.

Inherited Methods

`getReferencePoint`	Get the nth reference point.
`getReferenceImagePointsList`	Return the vector of reference image point.
`getMatchedPointNumber`	Get the number of matched points.
`getMatchedReferencePoints`	Return the index of the matched associated to the current image point i.
`referenceBuilt`	Indicate wether the reference has been built or not.
`getReferencePointNumber`	Get the number of reference points.
`getCurrentImagePointsList`	Return the vector of current image point.
`getIndexInAllReferencePointList`	Get the nth matched reference point index in the complete list of reference point.
`getMatchedPoints`	Get the nth couple of reference point and current point which have been matched.

Operators

`__doc__`
`__init__`	Overloaded function.
`__module__`

Attributes

`DESCRIPTOR_AKAZE`
`DESCRIPTOR_BRISK`
`DESCRIPTOR_KAZE`
`DESCRIPTOR_ORB`
`DESCRIPTOR_SIFT`
`DESCRIPTOR_TYPE_SIZE`
`DETECTOR_AGAST`
`DETECTOR_AKAZE`
`DETECTOR_BRISK`
`DETECTOR_FAST`
`DETECTOR_GFTT`
`DETECTOR_KAZE`
`DETECTOR_MSER`
`DETECTOR_ORB`
`DETECTOR_SIFT`
`DETECTOR_SimpleBlob`
`DETECTOR_TYPE_SIZE`
`__annotations__`
`constantFactorDistanceThreshold`
`detectionScore`
`detectionThreshold`
`jpgImageFormat`
`noFilterMatching`
`pgmImageFormat`
`pngImageFormat`
`ppmImageFormat`
`ratioDistanceThreshold`
`stdAndRatioDistanceThreshold`
`stdDistanceThreshold`

class DetectionMethodType(self, value: int)¶

Bases: pybind11_object

Predefined constant for descriptor extraction type.

Values:

DESCRIPTOR_ORB: ORB descriptor.
DESCRIPTOR_BRISK: BRISK descriptor.
DESCRIPTOR_SIFT: SIFT descriptor.
DESCRIPTOR_SURF: SUFT descriptor.
DESCRIPTOR_KAZE: KAZE descriptor.
DESCRIPTOR_AKAZE: AKAZE descriptor.
DESCRIPTOR_TYPE_SIZE: Number of descriptors available.

__and__(self, other: object) → object¶

__eq__(self, other: object) → bool¶

__ge__(self, other: object) → bool¶

__getstate__(self) → int¶

__gt__(self, other: object) → bool¶

__hash__(self) → int¶

__index__(self) → int¶

__init__(self, value: int)¶

__int__(self) → int¶

__invert__(self) → object¶

__le__(self, other: object) → bool¶

__lt__(self, other: object) → bool¶

__ne__(self, other: object) → bool¶

__or__(self, other: object) → object¶

__rand__(self, other: object) → object¶

__ror__(self, other: object) → object¶

__rxor__(self, other: object) → object¶

__setstate__(self, state: int) → None¶

__xor__(self, other: object) → object¶

property name : str¶

class FeatureDescriptorType(self, value: int)¶

Bases: pybind11_object

Predefined constant for descriptor extraction type.

Values:

DESCRIPTOR_ORB: ORB descriptor.
DESCRIPTOR_BRISK: BRISK descriptor.
DESCRIPTOR_SIFT: SIFT descriptor.
DESCRIPTOR_SURF: SUFT descriptor.
DESCRIPTOR_KAZE: KAZE descriptor.
DESCRIPTOR_AKAZE: AKAZE descriptor.
DESCRIPTOR_TYPE_SIZE: Number of descriptors available.

__and__(self, other: object) → object¶

__eq__(self, other: object) → bool¶

__ge__(self, other: object) → bool¶

__getstate__(self) → int¶

__gt__(self, other: object) → bool¶

__hash__(self) → int¶

__index__(self) → int¶

__init__(self, value: int)¶

__int__(self) → int¶

__invert__(self) → object¶

__le__(self, other: object) → bool¶

__lt__(self, other: object) → bool¶

__ne__(self, other: object) → bool¶

__or__(self, other: object) → object¶

__rand__(self, other: object) → object¶

__ror__(self, other: object) → object¶

__rxor__(self, other: object) → object¶

__setstate__(self, state: int) → None¶

__xor__(self, other: object) → object¶

property name : str¶

class FeatureDetectorType(self, value: int)¶

Bases: pybind11_object

Predefined constant for descriptor extraction type.

Values:

DESCRIPTOR_ORB: ORB descriptor.
DESCRIPTOR_BRISK: BRISK descriptor.
DESCRIPTOR_SIFT: SIFT descriptor.
DESCRIPTOR_SURF: SUFT descriptor.
DESCRIPTOR_KAZE: KAZE descriptor.
DESCRIPTOR_AKAZE: AKAZE descriptor.
DESCRIPTOR_TYPE_SIZE: Number of descriptors available.

__and__(self, other: object) → object¶

__eq__(self, other: object) → bool¶

__ge__(self, other: object) → bool¶

__getstate__(self) → int¶

__gt__(self, other: object) → bool¶

__hash__(self) → int¶

__index__(self) → int¶

__init__(self, value: int)¶

__int__(self) → int¶

__invert__(self) → object¶

__le__(self, other: object) → bool¶

__lt__(self, other: object) → bool¶

__ne__(self, other: object) → bool¶

__or__(self, other: object) → object¶

__rand__(self, other: object) → object¶

__ror__(self, other: object) → object¶

__rxor__(self, other: object) → object¶

__setstate__(self, state: int) → None¶

__xor__(self, other: object) → object¶

property name : str¶

class FilterMatchingType(self, value: int)¶

Bases: pybind11_object

Predefined constant for descriptor extraction type.

Values:

DESCRIPTOR_ORB: ORB descriptor.
DESCRIPTOR_BRISK: BRISK descriptor.
DESCRIPTOR_SIFT: SIFT descriptor.
DESCRIPTOR_SURF: SUFT descriptor.
DESCRIPTOR_KAZE: KAZE descriptor.
DESCRIPTOR_AKAZE: AKAZE descriptor.
DESCRIPTOR_TYPE_SIZE: Number of descriptors available.

__and__(self, other: object) → object¶

__eq__(self, other: object) → bool¶

__ge__(self, other: object) → bool¶

__getstate__(self) → int¶

__gt__(self, other: object) → bool¶

__hash__(self) → int¶

__index__(self) → int¶

__init__(self, value: int)¶

__int__(self) → int¶

__invert__(self) → object¶

__le__(self, other: object) → bool¶

__lt__(self, other: object) → bool¶

__ne__(self, other: object) → bool¶

__or__(self, other: object) → object¶

__rand__(self, other: object) → object¶

__ror__(self, other: object) → object¶

__rxor__(self, other: object) → object¶

__setstate__(self, state: int) → None¶

__xor__(self, other: object) → object¶

property name : str¶

class ImageFormatType(self, value: int)¶

Bases: pybind11_object

Predefined constant for descriptor extraction type.

Values:

DESCRIPTOR_ORB: ORB descriptor.
DESCRIPTOR_BRISK: BRISK descriptor.
DESCRIPTOR_SIFT: SIFT descriptor.
DESCRIPTOR_SURF: SUFT descriptor.
DESCRIPTOR_KAZE: KAZE descriptor.
DESCRIPTOR_AKAZE: AKAZE descriptor.
DESCRIPTOR_TYPE_SIZE: Number of descriptors available.

__and__(self, other: object) → object¶

__eq__(self, other: object) → bool¶

__ge__(self, other: object) → bool¶

__getstate__(self) → int¶

__gt__(self, other: object) → bool¶

__hash__(self) → int¶

__index__(self) → int¶

__init__(self, value: int)¶

__int__(self) → int¶

__invert__(self) → object¶

__le__(self, other: object) → bool¶

__lt__(self, other: object) → bool¶

__ne__(self, other: object) → bool¶

__or__(self, other: object) → object¶

__rand__(self, other: object) → object¶

__ror__(self, other: object) → object¶

__rxor__(self, other: object) → object¶

__setstate__(self, state: int) → None¶

__xor__(self, other: object) → object¶

property name : str¶

__init__(*args, **kwargs)¶

Overloaded function.

__init__(self: visp._visp.vision.KeyPoint, detectorType: visp._visp.vision.KeyPoint.FeatureDetectorType, descriptorType: visp._visp.vision.KeyPoint.FeatureDescriptorType, matcherName: str, filterType: visp._visp.vision.KeyPoint.FilterMatchingType = ratioDistanceThreshold) -> None
__init__(self: visp._visp.vision.KeyPoint, detectorName: str = ORB, extractorName: str = ORB, matcherName: str = BruteForce-Hamming, filterType: visp._visp.vision.KeyPoint.FilterMatchingType = ratioDistanceThreshold) -> None
__init__(self: visp._visp.vision.KeyPoint, detectorNames: list[str], extractorNames: list[str], matcherName: str = BruteForce, filterType: visp._visp.vision.KeyPoint.FilterMatchingType = ratioDistanceThreshold) -> None

buildReference(*args, **kwargs)¶

Overloaded function.

buildReference(self: visp._visp.vision.KeyPoint, I: visp._visp.core.ImageGray) -> int

Build the reference keypoints list.

Parameters:

I: Input reference image.

Returns:

The number of detected keypoints in the image I .

buildReference(self: visp._visp.vision.KeyPoint, I: visp._visp.core.ImageGray, iP: visp._visp.core.ImagePoint, height: int, width: int) -> int

Build the reference keypoints list in a region of interest in the image.

Parameters:

I: Input reference image.
iP: Position of the top-left corner of the region of interest.
height: Height of the region of interest.
width: Width of the region of interest.

Returns:

The number of detected keypoints in the current image I.

buildReference(self: visp._visp.vision.KeyPoint, I: visp._visp.core.ImageGray, rectangle: visp._visp.core.Rect) -> int

Build the reference keypoints list in a region of interest in the image.

Parameters:

I: Input image.
rectangle: Rectangle of the region of interest.

Returns:

The number of detected keypoints in the current image I.

buildReference(self: visp._visp.vision.KeyPoint, I: visp._visp.core.ImageGray, trainKeyPoints: list[cv::KeyPoint], points3f: list[cv::Point3_<float>], append: bool = false, class_id: int = -1) -> tuple[int, list[cv::KeyPoint], list[cv::Point3_<float>]]

Build the reference keypoints list and compute the 3D position corresponding of the keypoints locations.

Parameters:

I: Input image.
trainKeyPoints: List of the train keypoints.
points3f: Output list of the 3D position corresponding of the keypoints locations.
append: If true, append the supply train keypoints with those already present.
class_id: The class id to be set to the input cv::KeyPoint if != -1.

Returns:

A tuple containing:

The number of detected keypoints in the current image I.
trainKeyPoints: List of the train keypoints.
points3f: Output list of the 3D position corresponding of the keypoints locations.

buildReference(self: visp._visp.vision.KeyPoint, I: visp._visp.core.ImageGray, trainKeyPoints: list[cv::KeyPoint], trainDescriptors: cv::Mat, points3f: list[cv::Point3_<float>], append: bool = false, class_id: int = -1) -> int

Build the reference keypoints list and compute the 3D position corresponding of the keypoints locations.

Parameters:

I: Input image.
trainKeyPoints: List of the train keypoints.
trainDescriptors: List of the train descriptors.
points3f: List of the 3D position corresponding of the keypoints locations.
append: If true, append the supply train keypoints with those already present.
class_id: The class id to be set to the input cv::KeyPoint if != -1.

Returns:

The number of keypoints in the current image I.

buildReference(self: visp._visp.vision.KeyPoint, I_color: visp._visp.core.ImageRGBa) -> int

Build the reference keypoints list.

Parameters:

I_color: Input reference image.

Returns:

The number of detected keypoints in the image I .

buildReference(self: visp._visp.vision.KeyPoint, I_color: visp._visp.core.ImageRGBa, iP: visp._visp.core.ImagePoint, height: int, width: int) -> int

Build the reference keypoints list in a region of interest in the image.

Parameters:

I_color: Input reference image.
iP: Position of the top-left corner of the region of interest.
height: Height of the region of interest.
width: Width of the region of interest.

Returns:

The number of detected keypoints in the current image I.

buildReference(self: visp._visp.vision.KeyPoint, I_color: visp._visp.core.ImageRGBa, rectangle: visp._visp.core.Rect) -> int

Build the reference keypoints list in a region of interest in the image.

Parameters:

I_color: Input image.
rectangle: Rectangle of the region of interest.

Returns:

The number of detected keypoints in the current image I.

buildReference(self: visp._visp.vision.KeyPoint, I_color: visp._visp.core.ImageRGBa, trainKeyPoints: list[cv::KeyPoint], points3f: list[cv::Point3_<float>], append: bool = false, class_id: int = -1) -> tuple[int, list[cv::KeyPoint], list[cv::Point3_<float>]]

Build the reference keypoints list and compute the 3D position corresponding of the keypoints locations.

Parameters:

I_color: Input image.
trainKeyPoints: List of the train keypoints.
points3f: Output list of the 3D position corresponding of the keypoints locations.
append: If true, append the supply train keypoints with those already present.
class_id: The class id to be set to the input cv::KeyPoint if != -1.

Returns:

A tuple containing:

The number of detected keypoints in the current image I.
trainKeyPoints: List of the train keypoints.
points3f: Output list of the 3D position corresponding of the keypoints locations.

buildReference(self: visp._visp.vision.KeyPoint, I_color: visp._visp.core.ImageRGBa, trainKeyPoints: list[cv::KeyPoint], trainDescriptors: cv::Mat, points3f: list[cv::Point3_<float>], append: bool = false, class_id: int = -1) -> int

Build the reference keypoints list and compute the 3D position corresponding of the keypoints locations.

Parameters:

I_color: Input image.
trainKeyPoints: List of the train keypoints.
trainDescriptors: List of the train descriptors.
points3f: List of the 3D position corresponding of the keypoints locations.
append: If true, append the supply train keypoints with those already present.
class_id: The class id to be set to the input cv::KeyPoint if != -1.

Returns:

The number of detected keypoints in the current image I.

static compute3D(*args, **kwargs)¶

Overloaded function.

compute3D(candidate: cv::KeyPoint, roi: list[visp._visp.core.Point], cam: visp._visp.core.CameraParameters, cMo: visp._visp.core.HomogeneousMatrix, point: cv::Point3_<float>) -> None

Compute the 3D coordinate in the world/object frame given the 2D image coordinate and under the assumption that the point is located on a plane whose the plane equation is known in the camera frame. The Z-coordinate is retrieved according to the proportional relationship between the plane equation expressed in the normalized camera frame (derived from the image coordinate) and the same plane equation expressed in the camera frame.

Parameters:

candidate: Keypoint we want to compute the 3D coordinate.
roi: List of 3D points in the camera frame representing a planar face.
cam: Camera parameters.
cMo: Homogeneous matrix between the world and the camera frames.
point: 3D coordinate in the world/object frame computed.

compute3D(candidate: visp._visp.core.ImagePoint, roi: list[visp._visp.core.Point], cam: visp._visp.core.CameraParameters, cMo: visp._visp.core.HomogeneousMatrix, point: visp._visp.core.Point) -> None

Parameters:

candidate: vpImagePoint we want to compute the 3D coordinate.
roi: List of 3D points in the camera frame representing a planar face.
cam: Camera parameters.
cMo: Homogeneous matrix between the world and the camera frames.
point: 3D coordinate in the world/object frame computed.

createImageMatching(*args, **kwargs)¶

Overloaded function.

createImageMatching(self: visp._visp.vision.KeyPoint, IRef: visp._visp.core.ImageGray, ICurrent: visp._visp.core.ImageGray, IMatching: visp._visp.core.ImageGray) -> None

Initialize the size of the matching image (case with a matching side by side between IRef and ICurrent).

Parameters:

IRef: Reference image.
ICurrent: Current image.
IMatching: Image matching.

createImageMatching(self: visp._visp.vision.KeyPoint, ICurrent: visp._visp.core.ImageGray, IMatching: visp._visp.core.ImageGray) -> None

Initialize the size of the matching image with appropriate size according to the number of training images. Used to display the matching of keypoints detected in the current image with those detected in multiple training images.

Parameters:

ICurrent: Current image.
IMatching: Image initialized with appropriate size.

createImageMatching(self: visp._visp.vision.KeyPoint, IRef: visp._visp.core.ImageGray, ICurrent: visp._visp.core.ImageRGBa, IMatching: visp._visp.core.ImageRGBa) -> None

Initialize the size of the matching image (case with a matching side by side between IRef and ICurrent).

Parameters:

IRef: Reference image.
ICurrent: Current image.
IMatching: Image matching.

createImageMatching(self: visp._visp.vision.KeyPoint, ICurrent: visp._visp.core.ImageRGBa, IMatching: visp._visp.core.ImageRGBa) -> None

Parameters:

ICurrent: Current image.
IMatching: Image initialized with appropriate size.

detect(*args, **kwargs)¶

Overloaded function.

detect(self: visp._visp.vision.KeyPoint, I: visp._visp.core.ImageGray, keyPoints: list[cv::KeyPoint], rectangle: visp._visp.core.Rect = vpRect()) -> list[cv::KeyPoint]

Detect keypoints in the image.

Parameters:

I: Input image.
keyPoints: Output list of the detected keypoints.
rectangle: Optional rectangle of the region of interest.

Returns:

A tuple containing:

keyPoints: Output list of the detected keypoints.

detect(self: visp._visp.vision.KeyPoint, I_color: visp._visp.core.ImageRGBa, keyPoints: list[cv::KeyPoint], rectangle: visp._visp.core.Rect = vpRect()) -> list[cv::KeyPoint]

Detect keypoints in the image.

Parameters:

I_color: Input image.
keyPoints: Output list of the detected keypoints.
rectangle: Optional rectangle of the region of interest.

Returns:

A tuple containing:

keyPoints: Output list of the detected keypoints.

detect(self: visp._visp.vision.KeyPoint, matImg: cv::Mat, keyPoints: list[cv::KeyPoint], mask: cv::Mat) -> list[cv::KeyPoint]

Detect keypoints in the image.

Parameters:

matImg: Input image.
keyPoints: Output list of the detected keypoints.
mask: Optional 8-bit integer mask to detect only where mask[i][j] != 0.

Returns:

A tuple containing:

keyPoints: Output list of the detected keypoints.

detect(self: visp._visp.vision.KeyPoint, I: visp._visp.core.ImageGray, keyPoints: list[cv::KeyPoint], elapsedTime: float, rectangle: visp._visp.core.Rect = vpRect()) -> tuple[list[cv::KeyPoint], float]

Detect keypoints in the image.

Parameters:

I: Input image.
keyPoints: Output list of the detected keypoints.
elapsedTime: Elapsed time.
rectangle: Optional rectangle of the region of interest.

Returns:

A tuple containing:

keyPoints: Output list of the detected keypoints.
elapsedTime: Elapsed time.

detect(self: visp._visp.vision.KeyPoint, I_color: visp._visp.core.ImageRGBa, keyPoints: list[cv::KeyPoint], elapsedTime: float, rectangle: visp._visp.core.Rect = vpRect()) -> tuple[list[cv::KeyPoint], float]

Detect keypoints in the image.

Parameters:

I_color: Input image.
keyPoints: Output list of the detected keypoints.
elapsedTime: Elapsed time.
rectangle: Optional rectangle of the region of interest.

Returns:

A tuple containing:

keyPoints: Output list of the detected keypoints.
elapsedTime: Elapsed time.

detect(self: visp._visp.vision.KeyPoint, matImg: cv::Mat, keyPoints: list[cv::KeyPoint], elapsedTime: float, mask: cv::Mat) -> tuple[list[cv::KeyPoint], float]

Detect keypoints in the image.

Parameters:

matImg: Input image.
keyPoints: Output list of the detected keypoints.
elapsedTime: Elapsed time.
mask: Optional 8-bit integer mask to detect only where mask[i][j] != 0.

Returns:

A tuple containing:

keyPoints: Output list of the detected keypoints.
elapsedTime: Elapsed time.

display(*args, **kwargs)¶

Overloaded function.

display(self: visp._visp.vision.KeyPoint, IRef: visp._visp.core.ImageGray, ICurrent: visp._visp.core.ImageGray, size: int = 3) -> None

Display the reference and the detected keypoints in the images.

Parameters:

IRef: Input reference image.
ICurrent: Input current image.
size: Size of the displayed cross.

display(self: visp._visp.vision.KeyPoint, ICurrent: visp._visp.core.ImageGray, size: int = 3, color: visp._visp.core.Color = vpColor::green) -> None

Display the reference keypoints.

Parameters:

ICurrent: Input current image.
size: Size of the displayed crosses.
color: Color of the crosses.

display(self: visp._visp.vision.KeyPoint, IRef: visp._visp.core.ImageRGBa, ICurrent: visp._visp.core.ImageRGBa, size: int = 3) -> None

Display the reference and the detected keypoints in the images.

Parameters:

IRef: Input reference image.
ICurrent: Input current image.
size: Size of the displayed cross.

display(self: visp._visp.vision.KeyPoint, ICurrent: visp._visp.core.ImageRGBa, size: int = 3, color: visp._visp.core.Color = vpColor::green) -> None

Display the reference keypoints.

Parameters:

ICurrent: Input current image.
size: Size of the displayed crosses.
color: Color of the crosses.

displayMatching(*args, **kwargs)¶

Overloaded function.

displayMatching(self: visp._visp.vision.KeyPoint, IRef: visp._visp.core.ImageGray, IMatching: visp._visp.core.ImageGray, crossSize: int, lineThickness: int = 1, color: visp._visp.core.Color = vpColor::green) -> None

Display the matching lines between the detected keypoints with those detected in one training image.

Parameters:

IRef: Reference image, used to have the x-offset.
IMatching: Resulting image matching.
crossSize: Size of the displayed crosses.
lineThickness: Thickness of the displayed lines.
color: Color to use, if none, we pick randomly a color for each pair of matching.

displayMatching(self: visp._visp.vision.KeyPoint, ICurrent: visp._visp.core.ImageGray, IMatching: visp._visp.core.ImageGray, ransacInliers: list[visp._visp.core.ImagePoint] = [], crossSize: int = 3, lineThickness: int = 1) -> None

Display matching between keypoints detected in the current image and with those detected in the multiple training images. Display also RANSAC inliers if the list is supplied.

Parameters:

ICurrent: Current image.
IMatching: Resulting matching image.
ransacInliers: List of Ransac inliers or empty list if not available.
crossSize: Size of the displayed crosses.
lineThickness: Thickness of the displayed line.

displayMatching(self: visp._visp.vision.KeyPoint, IRef: visp._visp.core.ImageGray, IMatching: visp._visp.core.ImageRGBa, crossSize: int, lineThickness: int = 1, color: visp._visp.core.Color = vpColor::green) -> None

Display the matching lines between the detected keypoints with those detected in one training image.

Parameters:

IRef: Reference image, used to have the x-offset.
IMatching: Resulting image matching.
crossSize: Size of the displayed crosses.
lineThickness: Thickness of the displayed lines.
color: Color to use, if none, we pick randomly a color for each pair of matching.

displayMatching(self: visp._visp.vision.KeyPoint, IRef: visp._visp.core.ImageRGBa, IMatching: visp._visp.core.ImageRGBa, crossSize: int, lineThickness: int = 1, color: visp._visp.core.Color = vpColor::green) -> None

Display the matching lines between the detected keypoints with those detected in one training image.

Parameters:

IRef: Reference image, used to have the x-offset.
IMatching: Resulting image matching.
crossSize: Size of the displayed crosses.
lineThickness: Thickness of the displayed lines.
color: Color to use, if none, we pick randomly a color for each pair of matching.

displayMatching(self: visp._visp.vision.KeyPoint, ICurrent: visp._visp.core.ImageRGBa, IMatching: visp._visp.core.ImageRGBa, ransacInliers: list[visp._visp.core.ImagePoint] = [], crossSize: int = 3, lineThickness: int = 1) -> None

Display matching between keypoints detected in the current image and with those detected in the multiple training images. Display also RANSAC inliers if the list is supplied.

Parameters:

ICurrent: Current image.
IMatching: Resulting matching image.
ransacInliers: List of Ransac inliers or empty list if not available.
crossSize: Size of the displayed crosses.
lineThickness: Thickness of the displayed line.

getCovarianceMatrix(self) → visp._visp.core.Matrix¶: Get the covariance matrix when estimating the pose using the Virtual Visual Servoing approach.

Warning

The compute covariance flag has to be true if you want to compute the covariance matrix.

Note

See setCovarianceComputation

getCurrentImagePointsList(self) → list[visp._visp.core.ImagePoint]¶

Return the vector of current image point.

Returns:: Vector of the current image point.

getDetectionTime(self) → float¶

Get the elapsed time to compute the keypoint detection.

Returns:: The elapsed time.

getDetector(*args, **kwargs)¶

Overloaded function.

getDetector(self: visp._visp.vision.KeyPoint, type: visp._visp.vision.KeyPoint.FeatureDetectorType) -> cv::Ptr<cv::Feature2D>
getDetector(self: visp._visp.vision.KeyPoint, name: str) -> cv::Ptr<cv::Feature2D>

Parameters:

name: Name of the detector.

Returns:

The detector or nullptr if the name passed in parameter does not exist.

getDetectorNames(self) → dict[visp._visp.vision.KeyPoint.FeatureDetectorType, str]¶: Get the feature detector name associated to the type.

getExtractionTime(self) → float¶

Get the elapsed time to compute the keypoint extraction.

Returns:: The elapsed time.

getExtractor(*args, **kwargs)¶

Overloaded function.

getExtractor(self: visp._visp.vision.KeyPoint, type: visp._visp.vision.KeyPoint.FeatureDescriptorType) -> cv::Ptr<cv::Feature2D>
getExtractor(self: visp._visp.vision.KeyPoint, name: str) -> cv::Ptr<cv::Feature2D>

Parameters:

name: Name of the descriptor extractor.

Returns:

The descriptor extractor or nullptr if the name passed in parameter does not exist.

getExtractorNames(self) → dict[visp._visp.vision.KeyPoint.FeatureDescriptorType, str]¶: Get the feature descriptor extractor name associated to the type.

getImageFormat(self) → visp._visp.vision.KeyPoint.ImageFormatType¶

Get the image format to use when saving training images.

Returns:: The image format.

getIndexInAllReferencePointList(self, indexInMatchedPointList: int) → int¶

Get the nth matched reference point index in the complete list of reference point.

In the code below referencePoint1 and referencePoint2 correspond to the same matched reference point.

vpKeyPoint keypoint;

// Here the code to compute the reference points and the current points.
vpImagePoint referencePoint1;
vpImagePoint currentPoint;
keypoint.getMatchedPoints(1, referencePoint1, currentPoint);  //Get the first matched points

vpImagePoint referencePoint2;
const vpImagePoint* referencePointsList = keypoint.getAllPointsInReferenceImage();
// Get the first matched reference point index in the complete reference point list
int index = keypoint.getIndexInAllReferencePointList(1);
// Get the first matched reference point
referencePoint2 = referencePointsList[index];

getMatchQueryToTrainKeyPoints(self) → list[tuple[cv::KeyPoint, cv::KeyPoint]]¶

Get the list of pairs with the correspondence between the matched query and train keypoints.

Returns:: The list of pairs with the correspondence between the matched query and train keypoints.

getMatchedPointNumber(self) → int¶

Get the number of matched points.

Returns:: the number of matched points.

getMatchedPoints(self, index: int, referencePoint: visp._visp.core.ImagePoint, currentPoint: visp._visp.core.ImagePoint) → None¶

Get the nth couple of reference point and current point which have been matched. These points are copied in the vpImagePoint instances given in argument.

Parameters:

index: int¶: The index of the desired couple of reference point and current point. The index must be between 0 and the number of matched points - 1.
referencePoint: visp._visp.core.ImagePoint¶: The coordinates of the desired reference point are copied here.
currentPoint: visp._visp.core.ImagePoint¶: The coordinates of the desired current point are copied here.

getMatchedReferencePoints(self) → list[int]¶

Return the index of the matched associated to the current image point i. The ith element of the vector is the index of the reference image point matching with the current image point.

Returns:: The vector of matching index.

getMatcher(self) → cv::Ptr<cv::DescriptorMatcher>¶

Get the matcher pointer.

Returns:: The matcher pointer.

getMatches(self) → list[cv::DMatch]¶

Get the list of matches (correspondences between the indexes of the detected keypoints and the train keypoints).

Returns:: The list of matches.

getMatchingTime(self) → float¶

Get the elapsed time to compute the matching.

Returns:: The elapsed time.

getNbImages(self) → int¶

Get the number of train images.

Returns:: The number of train images.

getObjectPoints(*args, **kwargs)¶

Overloaded function.

getObjectPoints(self: visp._visp.vision.KeyPoint, objectPoints: list[cv::Point3_<float>]) -> list[cv::Point3_<float>]

Get the 3D coordinates of the object points matched (the corresponding 3D coordinates in the object frame of the keypoints detected in the current image after the matching).

Parameters:

objectPoints: List of 3D coordinates in the object frame.

Returns:

A tuple containing:

objectPoints: List of 3D coordinates in the object frame.

getObjectPoints(self: visp._visp.vision.KeyPoint, objectPoints: list[visp._visp.core.Point]) -> list[visp._visp.core.Point]

Get the 3D coordinates of the object points matched (the corresponding 3D coordinates in the object frame of the keypoints detected in the current image after the matching).

Parameters:

objectPoints: List of 3D coordinates in the object frame.

Returns:

A tuple containing:

objectPoints: List of 3D coordinates in the object frame.

getPoseTime(self) → float¶

Get the elapsed time to compute the pose.

Returns:: The elapsed time.

getQueryDescriptors(self) → cv::Mat¶

Get the descriptors matrix for the query keypoints.

Returns:: Matrix with descriptors values at each row for each query keypoints.

getQueryKeyPoints(*args, **kwargs)¶

Overloaded function.

getQueryKeyPoints(self: visp._visp.vision.KeyPoint, keyPoints: list[cv::KeyPoint], matches: bool = true) -> list[cv::KeyPoint]

Get the query keypoints list in OpenCV type.

Parameters:

keyPoints: List of query keypoints (or keypoints detected in the current image).
matches: If false return the list of all query keypoints extracted in the current image. If true, return only the query keypoints list that have matches.

Returns:

A tuple containing:

keyPoints: List of query keypoints (or keypoints detected in the current image).

getQueryKeyPoints(self: visp._visp.vision.KeyPoint, keyPoints: list[visp._visp.core.ImagePoint], matches: bool = true) -> list[visp._visp.core.ImagePoint]

Get the query keypoints list in ViSP type.

Parameters:

keyPoints: List of query keypoints (or keypoints detected in the current image).
matches: If false return the list of all query keypoints extracted in the current image. If true, return only the query keypoints list that have matches.

Returns:

A tuple containing:

keyPoints: List of query keypoints (or keypoints detected in the current image).

getRansacInliers(self) → list[visp._visp.core.ImagePoint]¶

Get the list of Ransac inliers.

Returns:: The list of Ransac inliers.

getRansacOutliers(self) → list[visp._visp.core.ImagePoint]¶

Get the list of Ransac outliers.

Returns:: The list of Ransac outliers.

getReferenceImagePointsList(self) → list[visp._visp.core.ImagePoint]¶

Return the vector of reference image point.

Returns:: Vector of reference image point.

getReferencePoint(self, index: int, referencePoint: visp._visp.core.ImagePoint) → None¶

Get the nth reference point. This point is copied in the vpImagePoint instance given in argument.

Parameters:

index: int¶: The index of the desired reference point. The index must be between 0 and the number of reference points - 1.
referencePoint: visp._visp.core.ImagePoint¶: The coordinates of the desired reference point are copied there.

getReferencePointNumber(self) → int¶

Get the number of reference points.

Returns:: the number of reference points.

getTrainDescriptors(self) → cv::Mat¶

Get the train descriptors matrix.

Returns:: : Matrix with descriptors values at each row for each train keypoints (or reference keypoints).

getTrainKeyPoints(*args, **kwargs)¶

Overloaded function.

getTrainKeyPoints(self: visp._visp.vision.KeyPoint, keyPoints: list[cv::KeyPoint]) -> list[cv::KeyPoint]

Get the train keypoints list in OpenCV type.

Parameters:

keyPoints: List of train keypoints (or reference keypoints).

Returns:

A tuple containing:

keyPoints: List of train keypoints (or reference keypoints).

getTrainKeyPoints(self: visp._visp.vision.KeyPoint, keyPoints: list[visp._visp.core.ImagePoint]) -> list[visp._visp.core.ImagePoint]

Get the train keypoints list in ViSP type.

Parameters:

keyPoints: List of train keypoints (or reference keypoints).

Returns:

A tuple containing:

keyPoints: List of train keypoints (or reference keypoints).

getTrainPoints(*args, **kwargs)¶

Overloaded function.

getTrainPoints(self: visp._visp.vision.KeyPoint, points: list[cv::Point3_<float>]) -> list[cv::Point3_<float>]

Get the train points (the 3D coordinates in the object frame) list in OpenCV type.

Parameters:

points: List of train points (or reference points).

Returns:

A tuple containing:

points: List of train points (or reference points).

getTrainPoints(self: visp._visp.vision.KeyPoint, points: list[visp._visp.core.Point]) -> list[visp._visp.core.Point]

Get the train points (the 3D coordinates in the object frame) list in ViSP type.

Parameters:

points: List of train points (or reference points).

Returns:

A tuple containing:

points: List of train points (or reference points).

initMatcher(self, matcherName: str) → None¶

Initialize a matcher based on its name.

Parameters:

matcherName: str¶: Name of the matcher (e.g BruteForce, FlannBased).

insertImageMatching(*args, **kwargs)¶

Overloaded function.

insertImageMatching(self: visp._visp.vision.KeyPoint, IRef: visp._visp.core.ImageGray, ICurrent: visp._visp.core.ImageGray, IMatching: visp._visp.core.ImageGray) -> None

Insert a reference image and a current image side-by-side.

Parameters:

IRef: Reference image.
ICurrent: Current image.
IMatching: Matching image for displaying all the matching between the query keypoints and those detected in the training images.

insertImageMatching(self: visp._visp.vision.KeyPoint, ICurrent: visp._visp.core.ImageGray, IMatching: visp._visp.core.ImageGray) -> None

Insert the different training images in the matching image.

Parameters:

ICurrent: Current image.
IMatching: Matching image for displaying all the matching between the query keypoints and those detected in the training images

insertImageMatching(self: visp._visp.vision.KeyPoint, IRef: visp._visp.core.ImageRGBa, ICurrent: visp._visp.core.ImageRGBa, IMatching: visp._visp.core.ImageRGBa) -> None

Insert a reference image and a current image side-by-side.

Parameters:

IRef: Reference image.
ICurrent: Current image.
IMatching: Matching image for displaying all the matching between the query keypoints and those detected in the training images.

insertImageMatching(self: visp._visp.vision.KeyPoint, ICurrent: visp._visp.core.ImageRGBa, IMatching: visp._visp.core.ImageRGBa) -> None

Insert the different training images in the matching image.

Parameters:

ICurrent: Current image.
IMatching: Matching image for displaying all the matching between the query keypoints and those detected in the training images

loadConfigFile(self, configFile: str) → None¶

Load configuration parameters from an XML config file.

Parameters:

configFile: str¶: Path to the XML config file.

loadLearningData(self, filename: str, binaryMode: bool = false, append: bool = false) → None¶

Load learning data saved on disk.

Parameters:

filename: str¶: Path of the learning file.
binaryMode: bool = false¶: If true, the learning file is in a binary mode, otherwise it is in XML mode.
append: bool = false¶: If true, concatenate the learning data, otherwise reset the variables.

match(self: visp._visp.vision.KeyPoint, trainDescriptors: cv::Mat, queryDescriptors: cv::Mat, matches: list[cv::DMatch], elapsedTime: float) → tuple[list[cv::DMatch], float]¶

Match keypoints based on distance between their descriptors.

Parameters:

trainDescriptors: Train descriptors (or reference descriptors).
queryDescriptors: Query descriptors.
matches: Output list of matches.
elapsedTime: Elapsed time.

Returns:

A tuple containing:

matches: Output list of matches.
elapsedTime: Elapsed time.

matchPoint(*args, **kwargs)¶

Overloaded function.

matchPoint(self: visp._visp.vision.KeyPoint, I: visp._visp.core.ImageGray) -> int

Match keypoints detected in the image with those built in the reference list.

Parameters:

I: Input current image.

Returns: