Development of Spectral-based Classification Schemes Using Unmanned Aerial System Imagery

Zarzar, C., Dash, P., Dyer, J., Hathcock, L. A., & Moorhead, R. J. (2015). Development of Spectral-based Classification Schemes Using Unmanned Aerial System Imagery. AAG Annual Meeting. Chicago, IL: AAG.

Unmanned Aerial Systems (UASs) provide the unique opportunity to gather data at high spatial and temporal resolution in areas where the conditions and/or costs for manned air flight would be impractical. The current project uses UAS imagery from the Lower Pearl River Watershed in Louisiana to develop spectral-based classification schemes for land cover and inundated area. The system flown is the Altavian Nova fitted with the IRIS imaging payload, which records imagery in three possible bands (red, green, blue, and/or near-infrared) at 1 inch pixels (at 400’ Above Ground Level (AGL)) over the study region. It is advantageous to use UAS imagery for the development of the spectral-based classification schemes because of the potential to classify features with a higher degree of precision and with a more frequent (and flexible) update cycle. These spectral-based classifications will provide information, including land-water separation, important to such applications as hydrology, ecology, and meteorology. Furthermore, because the UAS imagery is currently collected every two months over the same general area, the associated spectral-based classification schemes will allow for the implementation of change detection methods for analysis of land use/cover variability and potential surface water storage.