SAS: Software Application for SMASH

SAS: Software Application for SMASH (Spectral Mixture Analysis for Surveillance of Harmful Algal Blooms)

The Software Application for SMASH (Spectral Mixture Analysis for Surveillance of Harmful Algal Blooms), or SAS for short, is an application to facilitate mapping of potentially harmful algal blooms in reservoirs, rivers, and lakes from remotely sensed data. More specifically, SAS is designed to exploit the detailed observations of reflectance available within a hyperspectral image to infer which particular kinds of algae or cyanobacteria might be present within the water body. This type of spatially distributed taxonomic information is derived using a computational method known as Multiple Endmember Spectral Mixture Analysis. In addition to the hyperspectral image itself, the user must also provide a set of endmembers compiled within a spectral library. The endmembers essentially serve as pure, known examples of the taxa to be identified from the image. SAS then provides an integrated workflow for importing and subsetting an image, defining a region of interest, filtering the data spatially and spectrally, and calculating spectral indices. SAS also handles the import, resampling, and filtering of a spectral library and includes an interface for interactive viewing of library spectra. The Run SMASH portion of SAS enables the user to parameterize a SMASH run, assess the spectral separability of prospective endmembers, subset the library accordingly, and invoke the core MESMA algorithm. Finally, SAS allows for visualization and export of various data products including classified maps, taxonomic distributions, class masks and fraction images, customized false color composites, and a Root Mean Square Error (RMSE) image that summarizes the uncertainty associated with the spectral mixture models on a per-pixel basis.