Abstract

This study assesses the effectiveness of hyperspectral airborne imagery to map the intertidal vegetation at Roberts Bank, a large salt marsh and tidal flat complex in the Fraser River delta, British Columbia, Canada. This ecosystem is an important habitat for migratory and resident birds and fish and is part of the largest single salmon-producing stream in the world, the Fraser River. Imagery acquired by the compact airborne spectrographic imager (CASI) was radiometrically and geometrically corrected and mosaicked. A simple maximum likelihood classifier was used to classify intertidal vegetation cover. Training and test sites were collected using field observations and oblique aerial photographs. Seven vegetation-substrate classes were accurately (92.3%) classified, ranging from muddy sediments in the regularly inundated tidal flat and low marsh areas, to higher marsh communities of cattails, grasses, sedges, and rushes. This study showed that both the hyperspectral and hyperspatial characteristics of the CASI imagery were important for achieving high separability among the classes due to reduced spectral mixing of different substrates at the subpixel scale when compared with simulated Landsat and Ikonos imagery.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call