houghCircle

Name

houghCircle - Hough transform to detect circles

Syntax

houghCircle [<option>...] <inimage> <minradius> <maxradius>

Description

Performs a Hough transform to detect circles. The first band of the input image inimage is transformed to a three dimensional Hough space, a multi band image. The size of each output band will be identical to the size of the input band, and the number of bands will be maxradius-minradius+1.

The input image should be something like a thresholded edge image. All pixels larger than 0 are transformed to Hough space.

After creating the transform, the Hough space may be searched for local maxima. Within each band, only the largest local maximum within a wsxws area is registered. Besides, only maxima with number of updates above ul (on a scale 0.0 .. 1.0) are used. A list of these maxima may be written to stdout, depending on the options given.

Options

-list

Print a list of maxima to stdout. Default: Don't.

-tf transformFileName

Save the transform to transformFileName. Default: Don't save.

-rf reconstructionFileName

The Hough-space maxima are used to reconstruct the image, and the image is written to reconstructionFileName. Default: Don't reconstruct and save.

-tt transformTitle

Use transformTitle as title for transformed image. Default: "Hough Circle Transform".

-rt reconstructionTitle

Use reconstructionTitle as title for reconstructed image. Default: "Hough Circle Reconstruction".

-ul updateLimit

This determines the minimum number of updates for a maximum to be used. The minimum number is determined from updateLimit and the radius corresponding to the current band:

 updateLimit * 2.0 * Pi * radius

Default: 0.5.

-ws windowSize

Default: 7.

Restrictions

Input image must have pixel type unsigned byte.

Return value

0 when ok, positive otherwise.

Author

Tor Lønnestad, BLAB, Ifi, UiO

Examples

 houghCircle -ws 11 -rf circles.img \\
     -rt DetectedCircles input.img 5 25

Id

$Id: houghCircle.c,v 1.33 1997/01/10 16:02:06 svein Exp $