You are here: Start » Function Reference » Computer Vision » Image Analysis » DetectCorners_CornerResponse

DetectCorners_CornerResponse

Header:	AVL.h
Namespace:	avl
Module:	FoundationBasic

Detects corners using corner response method.

Applications: Detection of characteristic points on an image.

Syntax

void avl::DetectCorners_CornerResponse
(
	const avl::Image& inMonoImage,
	atl::Optional<const avl::Region&> inRoi,
	const avl::CornerResponseMethod::Type inCornerResponseMethod,
	const int inKernelSize,
	atl::Optional<float> inThreshold,
	atl::Optional<atl::Array<avl::Point2D>&> outCorners,
	atl::Optional<avl::Image&> outCornerResponseImage = atl::NIL
)

Parameters

Name	Type	Range	Default	Description
inMonoImage	const Image&			Input image
inRoi	Optional<const Region&>		NIL	Range of pixels to be processed
inCornerResponseMethod	const CornerResponseMethod::Type			Method for computing corner response
inKernelSize	const int	1 - 10	3	Method kernel size
inThreshold	Optional<float>	0.0 - 255.0	50.0f	Threshold for corner response value, between 0 and 255, default value is taken from SelectThresholdValue on outCornerResponseImage and entropy method
outCorners	Optional<Array<Point2D>&>			Found corner points
outCornerResponseImage	Optional<Image&>		NIL

Requirements

For input inMonoImage only pixel formats are supported: 1⨯uint8, 1⨯int8, 1⨯uint16, 1⨯int16, 1⨯int32, 1⨯real.

Read more about pixel formats in Image documentation.

Optional Outputs

The computation of following outputs can be switched off by passing value atl::NIL to these parameters: outCorners, outCornerResponseImage.

Description

The operation detects corners using either Harris or Kanade-Tomasi corner response method, depending on inCornerResponseMethod. For every square window of size inKernelSize a convolution matrix is computed: \[M = \left(\begin{array}{ccc} \sum g_r^2 & \sum g_c g_r \\ \sum g_c g_r & \sum g_c^2 \end{array} \right) \] where the summation is performed over the whole window and \(g_r, g_c\) denote horizontal and vertical gradient respectively at the point.
Harris' corner response is computed the following way: \[H = \mathrm{det}(M) - k \mathrm{tr}^2(M) \] where k is a constant set to 0.01.
Kanade-Tomasi corner response is given by: \[KT = \mathrm{min}(\lambda_1, \lambda_2)\] where \(\lambda_1, \lambda_2\) are eigenvalues of the convolution matrix.
Values H or KT give corner response image.

Then a few steps are performed in order to extract corner points. First, the normalized corner response function is thresholded with inThreshold with a small hysteresis, then the remained points are split into connected regions (blobs) and the center of each blob is determined.

Examples

DetectCorners_CornerResponse with inThreshold=50, methods Harris and Kanade-Tomasi respectively.

Remarks

Both methods give similar results and are quite fast. However, they tend to be less accurate than DetectCorners_Foerstner.

Errors

List of possible exceptions:

Error type	Description
DomainError	Unknown algorithm in DetectCorners_CornerResponse.
DomainError	Not supported inMonoImage pixel format in DetectCorners_CornerResponse. Supported formats: 1xUInt8, 1xInt8, 1xUInt16, 1xInt16, 1xInt32, 1xReal.