points_harris_binomial (Operator)

Name

points_harris_binomial — Detect points of interest using the binomial approximation of the Harris operator.

Synopsis

points_harris_binomial(Image : : MaskSizeGrd, MaskSizeSmooth, Alpha, Threshold, Subpix : Row, Column)

Description

points_harris extracts points of interest from an one-channel BYTE image Image. The Harris operator is based upon the smoothed matrix

        /                                   \
        |         2                         |
        |       Ix               Ix Iy      |
M = G * |                                   |,
        |                          2        |
        |       Ix Iy            Iy         |
        \                                   /

where G stands for a Binomial smoothing of size MaskSizeSmooth and Ix and Iy are the first derivatives of the image, computed with a Sobel filter of size MaskSizeGrd. The resulting points are the positive local extrema of

                         2
Det(M) - Alpha*(Trace(M)) .

If necessary, they can be restricted to points with a minimum filter response of Threshold. By default the coordinates of the points are calculated with subpixel accuracy. This can be turned off by setting the parameter Subpix to 'off'.

Parameters

Image (input_object)	(multichannel-)image → object (byte)
Input image.

MaskSizeGrd (input_control)	integer → (integer)
Amount of binomial smoothing used for the calculation of the gradient.
Default value: 5
Suggested values: 3, 5, 7, 9, 11, 15, 21, 31
Typical range of values: 3 ≤ MaskSizeGrd ≤ 50
Recommended increment: 2
Restriction: MaskSizeGrd > 0

MaskSizeSmooth (input_control)	integer → (integer)
Amount of smoothing used for the integration of the gradients.
Default value: 15
Suggested values: 3, 5, 7, 9, 11, 15, 21, 31
Typical range of values: 3 ≤ MaskSizeSmooth ≤ 50
Recommended increment: 2
Restriction: MaskSizeSmooth > 0

Alpha (input_control)	real → (real)
Weight of the squared trace of the squared gradient matrix.
Default value: 0.04
Suggested values: 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08
Typical range of values: 0.001 ≤ Alpha ≤ 0.1
Minimum increment: 0.001
Recommended increment: 0.01
Restriction: Alpha > 0.0

Threshold (input_control)	number → (real / integer)
Minimum filter response for the points.
Default value: 1000.0
Restriction: Threshold >= 0.0

Subpix (input_control)	string → (string)
Turn on or off subpixel refinement.
Default value: 'on'
List of values: 'on', 'off'

Row (output_control)	point.y-array → (real)
Row coordinates of the detected points.

Column (output_control)	point.x-array → (real)
Column coordinates of the detected points.

Result

points_harris returns 2 (H_MSG_TRUE) if all parameters are correct and no error occurs during the execution. If the input is empty the behavior can be set via set_system('no_object_result',<Result>). If necessary, an exception is raised.

Parallelization Information

points_harris_binomial is reentrant and processed without parallelization.

Possible Successors

gen_cross_contour_xld

Alternatives

points_foerstner, points_harris, points_lepetit, points_sojka

References

C. Harris, M. Stephens: “A combined corner and edge detector”. Proceedings of the 4th Alvey Vision Conference, pp. 147-151, 1988.
V. Gouet, N.Boujemaa: “Object-based queries using color points of interest”. IEEE Workshop on Content-Based Access of Image and Video Libraries, CVPR/CBAIVL 2001, Hawaii, USA, 2001.

Module

Foundation