作者: Chenghai Yang , J. H. Everitt
DOI: 10.1080/01431160903369626
关键词:
摘要: Broom snakeweed Gutierrezia sarothrae Pursh Britt. & Rusby is one of the most widespread and abundant rangeland weeds in western North America. The objectives this study were to evaluate airborne hyperspectral imagery compare it with aerial colour-infrared CIR photography multispectral digital for mapping broom infestations. Airborne along photographs images was acquired from a area south Texas. transformed using minimum noise fraction MNF then classified distance, Mahalanobis maximum likelihood, spectral angle mapper SAM classifiers. digitized respectively mosaicked as photographic image image; these same Accuracy assessment showed that likelihood classifier performed best three types images. overall accuracies three-class classification maps snakeweed, mixed woody herbaceous 91.0%, 92.5%, 95.0%, respectively, image, MNF-transformed image. Kappa analysis there no significant differences likelihood-based classifications among These results indicate can be used monitoring infestations on rangelands.