Off-campus South Dakota State University users: To download campus access theses, please use the following link to log into our proxy server with your South Dakota State University ID and password.

Non-South Dakota State University users: Please talk to your librarian about requesting this thesis through interlibrary loan.

Document Type

Thesis - University Access Only

Award Date

1991

Degree Name

Master of Science (MS)

Department / School

Electrical Engineering

First Advisor

Dennis Helder

Abstract

Banding, or scan-to-scan striping, is a noise pattern which corrupts image data acquired from the thematic mapper imaging systems aboard the Earth remote sensing satellites Landsats 4 and 5. This noise pattern appears as horizontal bands of alternating lighter and darker intensities which are approximately 17 pixels wide. The banding pattern is most apparent in relatively homogeneous or smooth image regions such as water or desert areas. Besides being visually disturbing, banding degrades the qualitative and quantitative analysis of the image data. A spatial filtering algorithm is developed to remove banding from geometrically corrected image data. A one-dimensional debanding filter is derived using a minimum mean square error approach. Since the filter is derived assuming stationary image and noise statistics, the spatial domain representation requires truncation of the filter in the vertical direction. A tolerance scheme is also incorporated which allows the filter to adapt to its environment in order to preserve edges. Performance of the algorithm is evaluated using simulated test images and actual satellite image data. The algorithm's performance is quantitatively measured on test images of stationary and nonstationary data by calculating a mean square error. Results of filtering actual image data are qualitatively evaluated and compared to other methods. The algorithm significantly reduces the banding noise and produces results as good as or better than other methods, with a faster computation time.

Library of Congress Subject Headings

Imaging systems -- Image quality
Landsat satellites
Earth resources technology satellites

Format

application/pdf

Number of Pages

108

Publisher

South Dakota State University

Share

COinS