| hough - hough transform for line detection |
[ht, rho_range] = hough(imbin) |
| imbin |
| binary image array (foreground equals 1) |
| ht |
| Hough transform accumulation matrix. Rho varies along rows; theta varies along columns. Theta range is 0-179 degrees w/ 180 samples. |
| rho_range |
| Vector containing the range of radius values. Equals -rmax:rmax, where rmax is the maximum radius possibly found in the input image. |
| Function hough calculates the hough transform of a binary image. The coordinate system is centered in the image, and the Y axis points downwards. Theta grows from X axis to Y axis. Negatives rho's point to the upper half of the image. |
initial_dir = PWD;
chdir (SIPDIR + 'images');
// ======= Example 1
im = imread('star.bmp');
im = bwborder(im);
xbasc()
imshow(im,2);
h = hough(im);
xbasc()
imshow(h,[]); // theta varies horizontally from 0 to 180
ht = 1*(h>= 40); // threshold the hough space
lim = ihough(ht,size(im)); // draw the detected lines
xbasc()
imshow(lim + 2*im +1, hotcolormap(4)) // detected lines shown in yellow
// ======= Example 2: how to obtain the parameters
//
// creating a empty picture with a line at y = -90
e = zeros(200,200);
e(10,:) = 1;
// (remember that the Y axis points downwards and is centered in the
// middle of the image)
// getting its hough transform, and finding the points
// corresponding to y=10
[h, rrange] = hough(e);
[r,c] = find(h == max(h))
// Gets the parameters of the line
theta = c - 1 // 90 degrees
rho = rrange(r) // -90 rho (upper half of image)
// thx to Herve Lombaert for inspiring example #2 !
chdir(initial_dir);
|
| There are no parameters for controlling the sampling of the parameter space. Also, the routine is too slow even for medium-sized images. It's purpose is for prototyping. When the function gets better, it will be transcribed to C language. |
| "Shape Analysis and Classification", L. da F. Costa and R. M. Cesar Jr., CRC Press. |
| "Practical Computer Vision using C", J. R. Parker, Wiley. |
| Ricardo Fabbri <rfabbri@if.sc.usp.br> |
| The latest version of the Scilab Image Processing toolbox can be found at |
| http://siptoolbox.sourceforge.net |
| ihough, drawline, edge |