Re: Oriented Hough Transform
- From: "Mittal" <mittalgbhatt@xxxxxxxxx>
- Date: 25 Aug 2005 14:09:34 -0700
Bob,
The paper I am trying to follow is using the oriented Hough transform
and given the complexity and processing time of the rest of the
algorithm of finding the 'true' edges/lines that contain 'shutter
edges', I think it will be beneficial to have as fewer lines as I can
have. But simply increasing the threshold to limit the number of lines
obtained from HT will not help since I might end up losing lines that
are part of the actual shutter border.
So if I take into account the orientation of the edges, I might be able
to retain more 'true shutter edges/lines' that probably contain more
edge-points having same edge direction; rather than those lines that
are retained simply due to more edge-points (that have very different
orientations and hence dont belong to the same edge) along the length
of that line.
I dont know if my reasoning is making enough sense, but thanks for all
your help and suggestions. I am doing some experimental coding to check
out the results of the algorithm and hence am using matlab, but my
final code needs to be in c++ so I cant rely too much on matlab inbuilt
functions. But I think that book is a good reference and I will look it
up nonetheless. Thanks a lot.
And if the edge-orientation accuracy can not be improved I probably
will go for the simple HT in the end.
Mittal
.
- References:
- Oriented Hough Transform
- From: Mittal
- Re: Oriented Hough Transform
- From: Bob Alvarez
- Re: Oriented Hough Transform
- From: Mittal
- Re: Oriented Hough Transform
- From: Bob Alvarez
- Oriented Hough Transform
- Prev by Date: Re: How do I analyse Fourier Transform?
- Next by Date: Measurement of Chromatic Aberration
- Previous by thread: Re: Oriented Hough Transform
- Next by thread: Video analysis library
- Index(es):
Relevant Pages
|