Clustering

I have an image with lot of line segments. I have to find the lines which have nearly same theta and rho (theta -orientation rho -the shortest distance between the line segment and the origin - like as in Hough transform). I have to find the lines which have nearly same theta and rho. How Can I do that?

Is it possible to do this by clustering rho theta matrix (Nx2 matrix)? if so what is the best algorithm to cluster this? (I don't know the number of clusters or centroids). I have tried with kmeans but didnot get good results.

Answer 1

[Elad]

try imopen with strel('line',rho,theta)