A new technique for performing temporal image normalization using pseudo-invariant features was investigated. The technique was applied to the six reflected spectral band images of the Landsat TM sensors. The temporal normalization of pseudo-invariant features yielded linear normalizing functions for all bands. The errors in the normalization of pseudo-invariant features was determined to be on the order of three digital counts, which was estimated to be equivalent to reflectance errors on the order of one percent reflectance. Temporal normalization of all features in the Landsat scene shows great potential for both quantitative and qualitative temporal change detection.
Library of Congress Subject Headings
Department, Program, or Center
Chester F. Carlson Center for Imaging Science (COS)
Volchok, William, "A Study of mutispectral temporal scene normalization using pseudo-invariant features, applied to Landsat TM imagery" (1985). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus