corner-foerstner

corner_foerstner

skimage.feature.corner_foerstner(image, sigma=1) [source]

Compute Foerstner corner measure response image.

This corner detector uses information from the auto-correlation matrix A:

1 2	A = [(imx**2)   (imx*imy)] = [Axx Axy]     [(imx*imy)   (imy**2)]   [Axy Ayy]

Where imx and imy are first derivatives, averaged with a gaussian filter. The corner measure is then defined as:

1 2	w = det(A) / trace(A)           (size of error ellipse) q = 4 * det(A) / trace(A)**2    (roundness of error ellipse)

Parameters:	image : ndarray Input image. sigma : float, optional Standard deviation used for the Gaussian kernel, which is used as weighting function for the auto-correlation matrix.
Returns:	w : ndarray Error ellipse sizes. q : ndarray Roundness of error ellipse.

References

[R132]

http://www.ipb.uni-bonn.de/uploads/tx_ikgpublication/foerstner87.fast.pdf

[R133]

http://en.wikipedia.org/wiki/Corner_detection

Examples

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

>>> from skimage.feature import corner_foerstner, corner_peaks >>> square = np.zeros([10, 10]) >>> square[2:8, 2:8] = 1 >>> square.astype(int) array([[0, 0, 0, 0, 0, 0, 0, 0, 0, 0],        [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],        [0, 0, 1, 1, 1, 1, 1, 1, 0, 0],        [0, 0, 1, 1, 1, 1, 1, 1, 0, 0],        [0, 0, 1, 1, 1, 1, 1, 1, 0, 0],        [0, 0, 1, 1, 1, 1, 1, 1, 0, 0],        [0, 0, 1, 1, 1, 1, 1, 1, 0, 0],        [0, 0, 1, 1, 1, 1, 1, 1, 0, 0],        [0, 0, 0, 0, 0, 0, 0, 0, 0, 0],        [0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]) >>> w, q = corner_foerstner(square) >>> accuracy_thresh = 0.5 >>> roundness_thresh = 0.3 >>> foerstner = (q > roundness_thresh) * (w > accuracy_thresh) * w >>> corner_peaks(foerstner, min_distance=1) array([[2, 2],        [2, 7],        [7, 2],        [7, 7]])