Remote sensing of Phragmites australis with the EO-1 Hyperion sensor

Phragmites australis is an invasive wetland plant that is recognized as a cause and a symptom of wetland degradation. This study evaluates the utility of Hyperion hyperspectral remote sensing imagery and common image analysis software for detecting large monodominant stands of Phragmites in coastal wetlands. Two approaches to hyperspectral image classification---unsupervised classification and target detection---are evaluated. The target detection approach achieved 68.3 percent overall accuracy with 41.2 percent user's accuracy. These results suggests that with further refinement of analysis techniques and the evolution of sensor technology, Hyperion and other space platform hyperspectral sensors may provide wetland scientists and resource managers with an efficient and effective monitoring tool.