Supervised classification of Landsat thematic mapper imagery in a semi-arid rangeland by nonparametric discriminant analysis
In this article the authors used a nonparametric discriminant function in a supervised classification of Landsat Thematic Mapper satellite imagery of a ~240,000-ha semi-arid region in the Snake River Plains, southwestern Idaho. First, agriculture pixels were classified by distance from the soil baseline and water pixels by the thermal band value. Next, successive nonparametric discriminant functions were used to separate grassland and shrubland categories with subsequent classifications of vegetation within major classes. Accuracy in separating grasslands and shrublands was 80 percent and remained consistent relative to different thresholds in minimum percent ground cover defining shrublands. Within major grassland shrubland groups, researchers achieved 64 percent accuracy in separating dominant vegetation classes.
Additional publication details
|Publication Subtype||Journal Article|
|Title||Supervised classification of Landsat thematic mapper imagery in a semi-arid rangeland by nonparametric discriminant analysis|
|Series title||Photogrammetric Engineering and Remote Sensing|
|Publisher||American Society for Photogrammetry and Remote Sensing|
|Contributing office(s)||Forest and Rangeland Ecosystem Science Center|