Abstract
Since 2004, airborne hyperspectral imagery has been acquired over the Sacramento - San Joaquin Delta in northern California to map submerged and floating invasive species and study how they affect the Delta ecosystem. Acquiring imagery over 2220 square kilometers of the Delta typically required 60-70 flightlines each year, which were then further processed to surface reflectance, georegistered, and prepared for analysis. Further, each flightline was processed using multiple spectral mapping methods such as spectral angle mapper, spectral mixture analysis, spectral indexes, and continuum removal over water and cellulose absorption bands. The outputs of these transformations were used as inputs to a Random Forests classifier. Concurrent with image acquisition, field data (800-2000 points) were collected across the Delta for training and validation of the classification products. The field data were divided into test and training polygons. These polygons were overlaid on the transformed files and pixel data were extracted corresponding to the polygons. The training data were used to train the Random Forests classifier to identify 10 classes (water, submerged aquatic vegetation, emergent marsh, soil, non-photosynthetic vegetation, water hyacinth, water primrose, pennywort, shadow, riparian vegetation). The classifier was validated quantitatively using the test data at both pixel and polygon level using overall accuracy and kappa metrics. The classifier was then applied to all flightlines and class maps were produced. Mosaics of these class maps are published in this dataset.
