CircleHoughTransform¶

class CircleHoughTransform(*args, **kwargs)¶

Bases: pybind11_object

Class that permits to detect 2D circles in a image using the gradient-based Circle Hough transform. Please find more information on the algorithm here

Overloaded function.

__init__(self: visp._visp.imgproc.CircleHoughTransform) -> None

Construct a new vpCircleHoughTransform object with default parameters.

__init__(self: visp._visp.imgproc.CircleHoughTransform, algoParams: visp._visp.imgproc.CircleHoughTransform.CircleHoughTransformParams) -> None
__init__(self: visp._visp.imgproc.CircleHoughTransform, jsonPath: str) -> None

Methods

`__init__`	Overloaded function.
`computeVotingMask`	Compute the mask containing pixels that voted for the detections .
`detect`	Overloaded function.
`getCannyThreshold`	Get internal Canny filter upper threshold.
`getCenterCandidatesList`	Get the list of Center Candidates, stored as pair <idRow, idCol>
`getCenterCandidatesVotes`	Get the number of votes of each Center Candidates.
`getCircleCandidates`	Get the Circle Candidates before merging step.
`getCircleCandidatesProbabilities`	Get the probabilities of the Circle Candidates.
`getCircleCandidatesVotes`	Get the votes of the circle candidates.
`getCircleCenterMinDist`	Get circles center min distance in pixels.
`getCircleMaxRadius`	Get circles max radius in pixels.
`getCircleMinRadius`	Get circles min radius in pixels.
`getDetectionsProbabilities`	Get the probabilities of the detections that are outputed by vpCircleHoughTransform::detect()
`getDetectionsVotes`	Get the number of votes for the detections that are outputed by vpCircleHoughTransform::detect()
`getDetectionsVotingPoints`	Get the points that voted for the detections that are outputed by vpCircleHoughTransform::detect() .
`getEdgeMap`	Get the Edge Map computed thanks to the Canny edge filter.
`getGradientX`	Get the gradient along the horizontal axis of the image.
`getGradientY`	Get the gradient along the vertical axis of the image.
`getRecordVotingPoints`	Returns true if it was asked to record the points that voted for the detections.
`init`
`initFromJSON`	Initialize all the algorithm parameters using the JSON file whose path is jsonPath .
`saveConfigurationInJSON`	Save the configuration of the detector in a JSON file described by the path jsonPath .
`setCannyBackend`	Set the backend to use to perform the Canny edge detection.
`setCannyThreshold`	Set the threshold for the Canny operator.
`setCannyThresholdRatio`	Set the Canny thresholds ratio that are used to automatically compute the Canny thresholds in case the user asks to.
`setCenterComputationParameters`	Set the parameters of the computation of the circle center candidates.
`setCircleCenterBoundingBox`	Set circles center min and max location in the image.
`setCircleCenterMinDist`	Set circles center min distance.
`setCircleMaxRadius`	Set circles max radius.
`setCircleMinRadius`	Set circles min radius.
`setCirclePerfectness`	Set circles perfectness, which corresponds to the threshold of the colinearity between the gradient of a point and the radius it would form with a center candidate to be able to vote.
`setFilteringAndGradientType`	Permits to choose the filtering + gradient operators to use.
`setGaussianParameters`	Set the parameters of the Gaussian filter, that permits to blur the gradients of the image.
`setGradientFilterAperture`	Set the parameters of the gradient filter (Sobel or Scharr) kernel size filters.
`setMask`	Set the mask that permits to ignore some pixels when performing the circle detection.
`setRadiusMergingThresholds`	Set the radius merging threshold used during the merging step in order to merge the circles that are similar.
`setRadiusRatioThreshold`	Set the parameters of the computation of the circle radius candidates.
`setRecordVotingPoints`	Permits to either activate or deactivate the memorization of the points that voted for the detected circles.
`toString`	Create a string with all Hough transform parameters.

Inherited Methods

Operators

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

Attributes

`__annotations__`
`edgeMapOff`
`edgeMapOn`

__init__(*args, **kwargs)¶

Overloaded function.

__init__(self: visp._visp.imgproc.CircleHoughTransform) -> None

Construct a new vpCircleHoughTransform object with default parameters.

__init__(self: visp._visp.imgproc.CircleHoughTransform, algoParams: visp._visp.imgproc.CircleHoughTransform.CircleHoughTransformParams) -> None
__init__(self: visp._visp.imgproc.CircleHoughTransform, jsonPath: str) -> None

computeVotingMask(self: visp._visp.imgproc.CircleHoughTransform, I: visp._visp.core.ImageGray, detections: list[visp._visp.core.ImageCircle], mask: Optional[vpImage<bool>], opt_votingPoints: Optional[list[list[tuple[int, int]]]]) → None¶

Compute the mask containing pixels that voted for the detections .

Parameters:

I: The image for which we want to have the information.
detections: Vector containing the list of vpImageCircle for which we want to know the voting points.
mask: Optional mask where pixels to exclude have a value set to false.
opt_votingPoints: Optional vector of pairs of pixel coordinates that voted for the detections .

detect(*args, **kwargs)¶

Overloaded function.

detect(self: visp._visp.imgproc.CircleHoughTransform, cv_I: cv::Mat) -> list[visp._visp.core.ImageCircle]
detect(self: visp._visp.imgproc.CircleHoughTransform, I: visp._visp.core.ImageRGBa) -> list[visp._visp.core.ImageCircle]

Convert the input image in a gray-scale image and then perform Circle Hough Transform to detect the circles in it.

Parameters:

I: The input color image.

Returns:

std::vector<vpImageCircle> The list of 2D circles detected in the image.

detect(self: visp._visp.imgproc.CircleHoughTransform, I: visp._visp.core.ImageGray) -> list[visp._visp.core.ImageCircle]

Perform Circle Hough Transform to detect the circles in a gray-scale image.

Parameters:

I: The input gray scale image.

Returns:

std::vector<vpImageCircle> The list of 2D circles detected in the image.

detect(self: visp._visp.imgproc.CircleHoughTransform, I: visp._visp.core.ImageGray, nbCircles: int) -> list[visp._visp.core.ImageCircle]

Perform Circle Hough Transform to detect the circles in in a gray-scale image. Get only the nbCircles circles having the greatest number of votes.

Parameters:

I: The input gray scale image.
nbCircles: The number of circles we want to get. If negative, all the circles will be returned, sorted such as result[0] has the highest number of votes and result[end -1] the lowest.

Returns:

std::vector<vpImageCircle> The list of 2D circles with the most number of votes detected in the image.

getCannyThreshold(self) → float¶: Get internal Canny filter upper threshold. When value is equal to -1 (default), it means that the threshold is computed automatically.

getCenterCandidatesList(self) → list[tuple[float, float]]¶

Get the list of Center Candidates, stored as pair <idRow, idCol>

Returns:: std::vector<std::pair<float, float> > The list of Center Candidates, stored as pair <idRow, idCol>

getCenterCandidatesVotes(self) → list[int]¶

Get the number of votes of each Center Candidates.

Returns:: std::vector<int> The number of votes of each Center Candidates, ordered in the same way than m_centerCandidatesList .

getCircleCandidates(self) → list[visp._visp.core.ImageCircle]¶

Get the Circle Candidates before merging step.

Returns:: std::vector<vpImageCircle> The list of circle candidates that were obtained before the merging step.

getCircleCandidatesProbabilities(self) → list[float]¶

Get the probabilities of the Circle Candidates.

Returns:: std::vector<float> The votes accumulator.

getCircleCandidatesVotes(self) → list[int]¶

Get the votes of the circle candidates.

Returns:: std::vector<unsigned int> The votes of the circle candidates.

getCircleCenterMinDist(self) → float¶: Get circles center min distance in pixels.

getCircleMaxRadius(self) → float¶: Get circles max radius in pixels.

getCircleMinRadius(self) → float¶: Get circles min radius in pixels.

getDetectionsProbabilities(self) → list[float]¶: Get the probabilities of the detections that are outputed by vpCircleHoughTransform::detect()

getDetectionsVotes(self) → list[int]¶: Get the number of votes for the detections that are outputed by vpCircleHoughTransform::detect()

getDetectionsVotingPoints(self) → list[list[tuple[int, int]]]¶: Get the points that voted for the detections that are outputed by vpCircleHoughTransform::detect() .

getEdgeMap(self) → visp._visp.core.ImageGray¶

Get the Edge Map computed thanks to the Canny edge filter.

Returns:: vpImage<unsigned char> The edge map computed during the edge detection step.

getGradientX(self) → visp._visp.core.ImageFloat¶

Get the gradient along the horizontal axis of the image.

Returns:: vpImage<float> The gradient along the horizontal axis of the image.

getGradientY(self) → visp._visp.core.ImageFloat¶

Get the gradient along the vertical axis of the image.

Returns:: vpImage<float> The gradient along the vertical axis of the image.

getRecordVotingPoints(self) → bool¶: Returns true if it was asked to record the points that voted for the detections.

init(self, algoParams: visp._visp.imgproc.CircleHoughTransform.CircleHoughTransformParams) → None¶

initFromJSON(self, jsonPath: str) → None¶

Initialize all the algorithm parameters using the JSON file whose path is jsonPath . Throw a ** vpException ** error if the file does not exist.

Parameters:

jsonPath: str¶: The path towards the JSON configuration file.

saveConfigurationInJSON(self, jsonPath: str) → None¶

Save the configuration of the detector in a JSON file described by the path jsonPath . Throw a ** vpException ** is the file cannot be created.

Parameters:

jsonPath: str¶: The path towards the JSON output file.

setCannyBackend(self, type: visp._visp.core.ImageFilter.CannyBackendType) → None¶

Set the backend to use to perform the Canny edge detection.

Parameters:

type: visp._visp.core.ImageFilter.CannyBackendType¶: The backend that must be used.

setCannyThreshold(self, lowerCannyThreshold: float, upperCannyThreshold: float) → None¶

Set the threshold for the Canny operator. Only value greater than this value are marked as an edge. If negative, the threshold is automatically computed.

Parameters:

lowerCannyThreshold: float¶: Canny filter lower threshold. When set to -1 (default), compute automatically this threshold.
upperCannyThreshold: float¶: Canny filter upper threshold. When set to -1 (default), compute automatically this threshold.

setCannyThresholdRatio(self, lowerThreshRatio: float, upperThreshRatio: float) → None¶

Set the Canny thresholds ratio that are used to automatically compute the Canny thresholds in case the user asks to.

Note

See vpCircleHoughTransform::setCannyThreshold(const float&, const float&)

Parameters:

lowerThreshRatio: float¶: The ratio of the upper threshold the lower threshold will be equal to.
upperThreshRatio: float¶: The ratio of pixels that must have a gradient lower than the upper threshold.

setCenterComputationParameters(self, dilatationSize: int, centerThresh: float, averagingWindowSize: int = 5, expectedNbCenters: int = -1) → None¶

Set the parameters of the computation of the circle center candidates.

Parameters:

dilatationSize: int¶: Kernel size of the dilatation operation used to detect the maxima in the center accumulator.
centerThresh: float¶: Minimum number of votes a point must exceed to be considered as center candidate.
averagingWindowSize: int = 5¶: Size of the averaging window around the maximum number of votes to compute the center candidate such as it is the barycenter of the window. Must be odd.
expectedNbCenters: int = -1¶: Expected number of centers in the image. If the number is negative, all the centers are kept. Otherwise, maximum up to this number of centers are kept.

setCircleCenterBoundingBox(self, center_min_x: int, center_max_x: int, center_min_y: int, center_max_y: int) → None¶

Set circles center min and max location in the image. If one value is equal to std::numeric_limits<int>::min or std::numeric_limits<int>::max() , the algorithm will set it either to -maxRadius or +maxRadius depending on if it is the lower or upper limit that is missing.

Parameters:

center_min_x: int¶: Center min location on the horizontal axis, expressed in pixels.
center_max_x: int¶: Center max location on the horizontal axis, expressed in pixels.
center_min_y: int¶: Center min location on the vertical axis, expressed in pixels.
center_max_y: int¶: Center max location on the vertical axis, expressed in pixels.

setCircleCenterMinDist(self, center_min_dist: float) → None¶

Set circles center min distance. Change this value to detect circles with different distances to each other.

Parameters:

center_min_dist: float¶: Center min distance in pixels.

setCircleMaxRadius(self, circle_max_radius: float) → None¶

Set circles max radius.

Parameters:

circle_max_radius: float¶: Max radius in pixels.

setCircleMinRadius(self, circle_min_radius: float) → None¶

Set circles min radius.

Parameters:

circle_min_radius: float¶: Min radius in pixels.

setCirclePerfectness(self, circle_perfectness: float) → None¶

Set circles perfectness, which corresponds to the threshold of the colinearity between the gradient of a point and the radius it would form with a center candidate to be able to vote. The formula to get the equivalent angle is: angle = acos(circle_perfectness) .

Parameters:

circle_perfectness: float¶: Circle perfectness. Value between 0 and 1. A perfect circle has value 1.

setFilteringAndGradientType(self, type: visp._visp.core.ImageFilter.CannyFilteringAndGradientType) → None¶

Permits to choose the filtering + gradient operators to use.

Parameters:

type: visp._visp.core.ImageFilter.CannyFilteringAndGradientType¶: The type of filtering + gradient operators to use.

setGaussianParameters(self, kernelSize: int, stdev: float) → None¶

Set the parameters of the Gaussian filter, that permits to blur the gradients of the image.

Parameters:

kernelSize: int¶: The size of the Gaussian kernel. Must be an odd value.
stdev: float¶: The standard deviation of the Gaussian function.

setGradientFilterAperture(self, apertureSize: int) → None¶

Set the parameters of the gradient filter (Sobel or Scharr) kernel size filters.

Parameters:

apertureSize: int¶: The size of the gradient filters kernel. Must be an odd value.

setMask(self: visp._visp.imgproc.CircleHoughTransform, mask: vpImage<bool>) → None¶

Set the mask that permits to ignore some pixels when performing the circle detection.

Parameters:

mask: A boolean image where pixels set to true means that the pixel must be considered and set to false means that the pixel must be ignored.

setRadiusMergingThresholds(self, radiusDifferenceThresh: float) → None¶

Set the radius merging threshold used during the merging step in order to merge the circles that are similar.

Parameters:

radiusDifferenceThresh: float¶: Maximum radius difference between two circle candidates to consider merging them.

setRadiusRatioThreshold(self, radiusRatioThresh: float) → None¶

Set the parameters of the computation of the circle radius candidates.

Parameters:

radiusRatioThresh: float¶: Minimum number of votes per radian a radius candidate RC_ij of a center candidate CeC_i must have in order that the circle of center CeC_i and radius RC_ij must be considered as circle candidate.

setRecordVotingPoints(self, record: bool) → None¶

Permits to either activate or deactivate the memorization of the points that voted for the detected circles.

Parameters:

record: bool¶: True to activate the feature, false to deactivate it.

toString(self) → str¶: Create a string with all Hough transform parameters.