Vegetation and Terrain Effects on Digital Classification of LANDSAT Imagery.

摘要

LANDSAT scenes from the northern Chihuahuan Desert, south-central New Mexico and western Texas, taken in March and August 1975, were analyzed using conventional supervised digital image analysis techniques. The study objective was to evaluate th digitally classified LANDSAT image against known ground condition and determine those vegetative and terrain factors that could aid the manual interpretation of the classified image. The interpretation of the image and correct classification of most areas were greatly assisted by knowledge of plant community-landform relations, the species phenological and physiognomic characteristics, and the reflectance-cover relations for the soil and vegetation conditions in each scene. (Author)