We are currently working on the problem of robustly tracking a natural feature in a stream of images. The tracking is done based on a localized search for the maximum correlation between a template and the image, in the neighborhood of the expected location of the feature in the new image. We are evaluating the use of normalized correlation as well as sum of squared difference (SSD) for the correlation.
This visual feature tracker will later be part of a larger system which will use the estimated location of the feature to update the position estimate of a rover which has been watching the fixed feature move relative to the rover as the rover traverses a site.
|
|
|
|
 |
This page is maintained by Matthew Deans, a
robograd
in the Carnegie Mellon University School of Computer Science.
Comments? Questions? Mail me at deano@ri.cmu.edu Last Modified March 5, 1997 |
