Accession Number : ADP002025

Title :   Optimal LANDSAT Transforms for Forest Applications,


Personal Author(s) : Logan,T. L. ; Strahler,A. H.

Report Date : JUN 1982

Pagination or Media Count : 14

Abstract : In agricultural applications of remote sensing, linear transforms of Landsat data, such as those of Kauth and Thomas, are known to be highly effective both for data compression and enhancement of crop identification accuracies. Typically, such transforms are based on the time-trajectory of crop pixels through measurement space as the crop increasingly obscures the soil, matures, scenesces, and is harvested. In natural vegetation applications, temporal variations are less important-- life-form differences among vegetation types lead to distinctive signatures for natural vegetation types that are more or less distinctive, independent of season. However, the signatures of natural vegetation types are greatly influenced by their topographic position on the landscape, due to factors of differential illumination and complex bidirectional reflectance distribution functions. Thus, the question arises whether there are one or more transforms of Landsat data, beyond those already explored, that can accentuate the separability of natural vegetation classes in areas of diverse topographic relief. To answer this question, we investigated eleven transforms of four Landsat MSS channels. These transforms were evaluated for their ability to distinguish among thirteen classes of natural vegetation in a small area of the Klamath Mountains in northern California, USA.

Descriptors :   *Image processing, Oregon, Forests, Aerial photography, Multiband spectral reconnaissance, Transformations(Mathematics), Classification, Scientific satellites, Statistical analysis

Distribution Statement : APPROVED FOR PUBLIC RELEASE