I am using the following code on a given image:
imshow(image);
[out1,out2]=ginput(1);
I am getting a non-integer relative number.
(row,column)
from the output of ginput()
. estimateGeometricTransform()
) to move points in one image to a second image. should I use the pixel number or the regular output of ginput()
?ginput
gives you relative positions, so
round(ginput(1))
should get you to the centre of the pixel, which is what you'll probably want, as it's difficult to select matching features on a level finer than pixels. If you can select features on a level finer than pixels, you should of course use that, and do that for both images.