Soft morphological filtering of spectral images

Authors:
Sami Lakaniemi;Pekka Toivanen;Ville Kyrki
Affiliations:
Lappeenranta University of Technology, Department of Information Technology, Lappeenranta, Finland;Lappeenranta University of Technology, Department of Information Technology, Lappeenranta, Finland;Lappeenranta University of Technology, Department of Information Technology, Lappeenranta, Finland
Venue:
SCIA'03 Proceedings of the 13th Scandinavian conference on Image analysis
Year:
2003

Citing 3
Cited 0

Noise in Color Digital Images

MWSCAS '98 Proceedings of the 1998 Midwest Symposium on Systems and Circuits
Image Analysis and Mathematical Morphology

Image Analysis and Mathematical Morphology
Handbook of Image and Video Processing (Communications, Networking and Multimedia)

Handbook of Image and Video Processing (Communications, Networking and Multimedia)

Quantified Score

Hi-index	0.01

Visualization

Abstract

In this study, noise reduction methods for spectral images using soft morphological filters are presented. Three methods are introduced: filtering individual pixels in the direction of their spectra, two-dimensional componentwise filtering, and three-dimensional cubic filtering. The research involved removing impulsive noise, particularly salt and pepper and bit error noise. Filter performance was measured both quantitatively using mean absolute error, mean square error and signal to-noise ratio and visually by observing individual spectral channels and individual spectra of filtered images. The best results were achieved using componentwise filtering according to both numerical and visual criteria.