htl

Purpose

Hough transform for detection of lines in images.

Synopsis

[HT,rho,theta] = htl(IM).
[HT,rho,theta] = htl(IM,M).
[HT,rho,theta] = htl(IM,M,N).
[HT,rho,theta] = htl(IM,M,N,trace).

Description

From an image IM, computes the integration of the values of the image over all the lines. The lines are parametrized using polar coordinates. The origin of the coordinates is fixed at the center of the image, and theta is the angle between the vertical axis and the perpendicular (to the line) passing through the origin. Only the values of IM exceeding 5 % of the maximum are taken into account (to speed up the algorithm).

Name	Description	Default value
IM	image to be analyzed (size (Xmax,Ymax))
M	desired number of samples along the radial axis	Xmax
N	desired number of samples along the azimutal (angle) axis	Ymax
trace	if nonzero, the progression of the algorithm is shown	0
HT	output matrix (MxN matrix)
rho	sequence of samples along the radial axis
theta	sequence of samples along the azimutal axis

When called without output arguments, htl displays HT using mesh.

Example

The Wigner-Ville distribution of a linear frequency modulation is almost perfectly concentrated (in the discrete case) on a straight line in the time-frequency plane. Thus, applying the Hough transform on this image will produce a representation with a peak, whose coordinates give estimates of the linear frequency modulation parameters (initial frequency and sweep rate):

         N=64; t=(1:N); y=fmlin(N,0.1,0.3); 
         IM=tfrwv(y,t,N); imagesc(IM); pause(1); 
         htl(IM,N,N,1);

Reference

[1] H. Maître ``Un Panorama de la Transformation de Hough'', Traitement du Signal, Vol 2, No 4, pp. 305-317, 1985.

© GdR ISIS - Contact