1
$\begingroup$

i applyed hough transform in my image and this is the result of hough table, i must get the maxima of rho in two basic angles of theta , the one is around -65 : -60 and the other is round 30:35

enter image description here

How can i get them? I tried the below code about hough but the houghpeaks function returns all the maxima peaks in the hough table. I want to detect the two basic angles as you can see

[hough,theta,rho] = hough(negfillHoles,'Theta',-90:1:89,'RhoResolution',2);
lenHough=length(hough(:));
maxpeaks=lenHough; %10000;
peaks = houghpeaks(hough,maxpeaks,'threshold',20); %50
x=theta(peaks(:,2));
y=rho(peaks(:,1));
plot(x,y,'o','color','r');

I will really appreciate your help, it is a matter of life or death to find this answer for my home work

enter image description here

Improve this question
$\endgroup$
1
  • $\begingroup$ First idea: start with taking only the N largest peaks in your Hough plane. Second idea: marginalize over $\rho$ first, then detect peaks in $\theta$ domain (1-D peak serach). Use these to narrow down your search in the 2-D Hough plane. $\endgroup$
    – Florian
    May 21, 2019 at 7:34

1 Answer 1

Reset to default
1
$\begingroup$

Why don't you try to mask your data first. E.g. take the variance for every $\theta$ value, so that you get a $\rho$ variance information for each $\theta$ value. Variance should be high around your peask, since all the other parts are basically black or gray - only in the peak region there is a lot of changes. You could also play around with a auto-correlation of the data along that axis. Then, you could threshold the resulting vector so that you have a better estimate where the peaks are. Then apply the peak detection algorithm only in those areas of $\theta$.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.