I'm somewhat new to labview, and am working on a program which requires me to track objects in an image. From what I understand, pattern matching, geometric matching, and object tracking can all be used to track objects in an image. I do understand that pattern matching uses pixel intensity information to find objects (and is good for when the image is clear), and geometric matching uses boundary edges. However, I don't really understand why this should be used over object tracking in any situation, or visa versa. I would appreciate any clarification on the three methods.
Solved! Go to Solution.
All the 3 has different objective. I am not trying to replace the Vision concept manual. You might want to look there.
1. Pattern matching is very fast and accurate. But doesn't work well if the gray level in the image are changing. For example effects of illumination change. Pattern matching will search for the pattern over the whole image.
2. Geometric matching is slower. But it is robust for illumination change, rotation, scale changes. It might be little bit less precise. It will also search over the whole image.
3. Object tracking is made to follow object that was already found and is moving on a static background. The main feature that it is looking for is the motion. It will search for the object in a video while the object moves not by much between the image frames.
Thanks - Amit