Prev: blockproc takes much longer than blkproc
Next: speed up cell array operations
From: Kenneth on 4 Aug 2010 10:14
I asked a smilar question before, but did not get anywhere. I drew an example which can be viewed here: http://www.flickr.com/photos/29009236(a)N02/4860312546/. we have about 2000 pictures similar to this one, with the particle in different positions and rotations. Is it possible to determine the rotation angle of the particle using matlab? our current solution is a program where we eyeball the rotation, and manually input the scan direction, but for 2000+ pictures, this is not the greatest. If we can determine the angle, we can write a peice for the program that will automatically set the scan direction.
From: Sean on 4 Aug 2010 11:25
"Kenneth " <cybeastfalzar99(a)yahoo.com> wrote in message <i3bsjc$ndp$1(a)fred.mathworks.com>...
> I asked a smilar question before, but did not get anywhere. I drew an example which can be viewed here: http://www.flickr.com/photos/29009236(a)N02/4860312546/. we have about 2000 pictures similar to this one, with the particle in different positions and rotations. Is it possible to determine the rotation angle of the particle using matlab? our current solution is a program where we eyeball the rotation, and manually input the scan direction, but for 2000+ pictures, this is not the greatest. If we can determine the angle, we can write a peice for the program that will automatically set the scan direction.

Do you have the Image Processing Toolbox?
Are all of the images roughly the same shape?

If so, use bwconncomp() and regionprops() to get the orientation of the object. Though this orientation is meaningless because of the bend, it can be used as a reference to your desired orientation.
I.e:
-Figure out what the orientation of the object is when it's how you want it.
-Figure out what it is in its rotated state.
-The difference is your rotation.
From: Kenneth on 4 Aug 2010 11:36
"Sean " <sean.dewolski(a)nospamplease.umit.maine.edu> wrote in message <i3c0oh$oa1$1(a)fred.mathworks.com>...
> "Kenneth " <cybeastfalzar99(a)yahoo.com> wrote in message <i3bsjc$ndp$1(a)fred.mathworks.com>...
> > I asked a smilar question before, but did not get anywhere. I drew an example which can be viewed here: http://www.flickr.com/photos/29009236(a)N02/4860312546/. we have about 2000 pictures similar to this one, with the particle in different positions and rotations. Is it possible to determine the rotation angle of the particle using matlab? our current solution is a program where we eyeball the rotation, and manually input the scan direction, but for 2000+ pictures, this is not the greatest. If we can determine the angle, we can write a peice for the program that will automatically set the scan direction.
>
> Do you have the Image Processing Toolbox?
> Are all of the images roughly the same shape?
>
> If so, use bwconncomp() and regionprops() to get the orientation of the object. Though this orientation is meaningless because of the bend, it can be used as a reference to your desired orientation.
> I.e:
> -Figure out what the orientation of the object is when it's how you want it.
> -Figure out what it is in its rotated state.

they are all roughly the same, and they aren't drawn, the real ones are microscope pics. We are trying to find the center point, but need to determine angle rotated from an axis to find scan angle. Ill try those two functions.
> -The difference is your rotation.
From: Kenneth on 4 Aug 2010 12:53
"Kenneth " <cybeastfalzar99(a)yahoo.com> wrote in message <i3c1d4$6ar$1(a)fred.mathworks.com>...
> "Sean " <sean.dewolski(a)nospamplease.umit.maine.edu> wrote in message <i3c0oh$oa1$1(a)fred.mathworks.com>...
> > "Kenneth " <cybeastfalzar99(a)yahoo.com> wrote in message <i3bsjc$ndp$1(a)fred.mathworks.com>...
> > > I asked a smilar question before, but did not get anywhere. I drew an example which can be viewed here: http://www.flickr.com/photos/29009236(a)N02/4860312546/. we have about 2000 pictures similar to this one, with the particle in different positions and rotations. Is it possible to determine the rotation angle of the particle using matlab? our current solution is a program where we eyeball the rotation, and manually input the scan direction, but for 2000+ pictures, this is not the greatest. If we can determine the angle, we can write a peice for the program that will automatically set the scan direction.
> >
> > Do you have the Image Processing Toolbox?
> > Are all of the images roughly the same shape?
> >
> > If so, use bwconncomp() and regionprops() to get the orientation of the object. Though this orientation is meaningless because of the bend, it can be used as a reference to your desired orientation.
> > I.e:
> > -Figure out what the orientation of the object is when it's how you want it.
> > -Figure out what it is in its rotated state.
>

adding some info - Is it possible to use DCM to figure out the rotation angle of the object, if so, how would I got about doing it?
> they are all roughly the same, and they aren't drawn, the real ones are microscope pics. We are trying to find the center point, but need to determine angle rotated from an axis to find scan angle. Ill try those two functions.
> > -The difference is your rotation.
From: Kenneth on 4 Aug 2010 13:56
Ok, that got messed up, the question is, can i use DCM to determine the rotation of the object? if so, how?
 |  Next  |  Last
Pages: 1 2
Prev: blockproc takes much longer than blkproc
Next: speed up cell array operations