73.
We assessed aerial hyperspectral imagery with high spatial (1.5 m) and spectral (8.9 nm) resolutions for detecting and mapping the early invasion by
Solidago altissima of understory vegetation in moist tall grassland. Generalized linear models (GLMs) were constructed to predict
S. altissima occurrence using 1.5 m pixels from hyperspectral data collected during the spring when understory vegetation was directly observable from above. A data set of presence–absence derived from percentage cover data was used for the analyses. The values of the area under the receiver operating characteristic (ROC) curve (AUC) ranged from 0.77–0.87 in the validation data set. Three minimum noise fraction (MNF) bands differentiated
S. altissima in the best-performing model (selected based on Akaike's information criterion) for the occurrence of
S. altissima. The results suggest that the aerial hyperspectral images obtained during spring before the seasonal development of the grass canopy are useful for the early detection and mapping of
S. altissima invading moist tall grassland.
相似文献