Comparison of Image Endmember- and Object-Based Classification of Very-High-Spatial-Resolution Unmanned Aircraft System (UAS) Narrow-Band Images for Mapping Riparian Forests and Other Land Covers

Riparian forests are critical for carbon storage, biodiversity, and river water quality. There has been an increasing use of very-high-spatial-resolution (VHR) unmanned aircraft systems (UAS)-based remote sensing riparian forest mapping. However, improved forest/zone monitoring, restoration, management, enhanced understanding the accuracy different classification methods mapping other land covers at high thematic resolution is necessary. Research that compares efficacies endmember- object-based applied to VHR (e.g., UAS) images limited. Using Sequential Maximum Angle Convex Cone (SMACC) endmember extraction algorithm (EEA) jointly with Spectral Mapper (SAM) classifier, a separate multiresolution segmentation/object-based method, we map forests/land compare accuracies accrued via application these two approaches narrow-band, UAS orthoimages collected over reaches/riparian areas in Austria. We assess effect pixel size on accuracy, 7 20 cm pixels, evaluate performance across multiple dates. Our findings show markedly higher than those endmember-based approach, where former generally have overall >85%. Poor likely due very small sizes, as well large number classes, relatively bands used. Object-based this context provides effective monitoring management.