06-23-2005 09:34 AM
06-25-2005 12:42 PM - edited 06-25-2005 12:42 PM
Message Edited by Philip C. on 06-25-2005 12:43 PM
07-07-2005 07:49 AM
Thanx Philip C.
its working nice,
i would like to overlay the common part of the pictures so i can get one large image (panorama)so i can measure it.
could you help me with that?
07-07-2005 08:51 AM
07-08-2005 05:44 PM
Hi 2face,
Can you explain in a little more detail what you are looking for in the final image.
Thanks,
Nipun M
Applications Engineer
National Instruments
07-11-2005 03:29 AM
07-11-2005 07:19 AM
An iterative approach it to compare the left most column of the right picture with the right most column of the left picture. For each row compare the value of the pixel. This comparison is dependent on the color system you use, for example 24-bit RGB you could calculate the difference between RGB1 and RGB 2 = |R1-R2|+|G1-G2| + |B1-B2| which will yield a confidence value that the pixels reference the same area of an image. Sum all the overlap values and divide by the total number of pixels compared. You will have an average overlap difference. Now inclement the overlap columns by one and compare again, you should see that the minimum value will occur near the optimal overlap region. This example assumes that the images are aligned up in the vertical direction already (such is the case when using a tripod). This is how I would approach this problem, I know this is a simplified solution but it is easy to implement and doesn't require fancy pattern matching or complex algorithms. I hope it helps get you started.
Paul
07-11-2005 08:46 AM
07-14-2005 02:43 AM
07-26-2005 07:30 AM