Fourier-space processing has been shown above to offer a powerful way to remove periodic noise from images. It is also used to perform convolutions, such as Gaussian smoothing or high-pass sharpening filters, more efficiently than can be done with large kernels applied directly to the pixel values. Deconvolution in which out-of-focus images are restored was also shown above.
Another very useful Fourier-space technique is cross-correlation. This requires two images, one of a scene containing objects of interest that may be difficult to locate or count because of their appearance or because of a complex surroundings, including camouflage. The second image contains just a single object of the type being sought. Fourier transforms of the two images are multiplied together with a phase shift and the result re-transformed to the pixel domain, to produce bright spots where the objects were located. The top hat filter, introduced above, may then be used to locate the spots.
In the Cross Correlation interactive Java tutorial, although human recognition of the various letters is easy, counting each one is difficult because of the irregular layout. Using each letter as a target, performing cross correlation, and applying a top hat filter produces a result that makes the counting process simple.
A more realistic application appears in the Applying Cross Correlation Java tutorial, which shows an image of bubbles, with illumination from one side that produces bright and dark regions that make automatic thresholding to detect the features difficult. Cross correlation with an image of a single bubble locates the features, even though they vary somewhat in size and texture.
John C. Russ - Materials Science and Engineering Dept., North Carolina State University, Raleigh, North Carolina, 27695.
Matthew Parry-Hill and Michael W. Davidson - National High Magnetic Field Laboratory, 1800 East Paul Dirac Dr., The Florida State University, Tallahassee, Florida, 32310.
Questions or comments? Send us an email.
© 1998-2009 by Michael W. Davidson, John Russ, Olympus America Inc., and The Florida State University. All Rights Reserved. No images, graphics, scripts, or applets may be reproduced or used in any manner without permission from the copyright holders. Use of this website means you agree to all of the Legal Terms and Conditions set forth by the owners.
This website is maintained by our