1

I want to detect vanishing point of an entire image that contain a bunch of vanishing lines..

The task i have completed already described below:

1.Detecting edges of man-made scenes from images 2.From that using hough transform develope/find straight lines from that images 3.Using these straight lines to find VANISHING POINT of the entire image

The steps 1 and 2 are now completed. My current problem is to find the VANISHING POINT from result of step 2.And no idea about that. Anybody can help me to find how to find vanishing point from the step 2.

T.J. Crowder
  • 1,031,962
  • 187
  • 1,923
  • 1,875
Rahul R Nath
  • 11
  • 1
  • 3

2 Answers2

0

The vanishing point is the crossing of the vanishing lines. So in theory finding the crossing of any two lines is enough.

As you have many lines, you can use it to increase the accuracy and robustness of your result. What I would do as a first step is to find the point whose sum of square distances from all lines is minimal. This results in simple linear regression. Further improvement can be improving the robustness against outliers - if these are an issue in your case.

Ophir Gvirtzer
  • 604
  • 4
  • 8
  • I got many lines from step 2 i mentioned above. There is an algorithm "Levenberg-Marquardt" use for detecting vanishing point from using all these straight lines. But i searched a lot on internet.Cant find that algorithm, or that algorithmic related codes!!!! – Rahul R Nath Apr 03 '14 at 09:08
0

I found an another method is vailable to find out vanishing point. Method:"PIXEL ORIENTATION BASED METHOD" Any body can give MATLAB code of finding vanishing point?

Rahul R Nath
  • 11
  • 1
  • 3