Geospatial object-based image analysis (GEOBIA) and Endmember-based Land-Cover Classification of Narrow-Band Remote-Sensor Image Data
Topics:
Keywords: endmember, GEOBIA, land cover, UAS
Abstract Type: Poster Abstract
Authors:
Anthony M. Filippi, Texas A&M University
Inci Güneralp, Texas A&M University
Cesar R. Castillo, Texas A&M University
Andong Ma, Texas A&M University
Gernot Paulus, Carinthia University of Applied Sciences, Austria
Karl-Heinrich Anders, Carinthia University of Applied Sciences, Austria
,
,
,
,
Abstract
Research that compares classification efficacies of GEOBIA and endmember-based methods applied to very-high-spatial-resolution (VHR) images (e.g., unmanned aircraft systems (UAS) images) is limited. Using an endmember extraction algorithm jointly with an endmember-based classifier, and a separate multiresolution segmentation/object-based classification method, we map riparian land covers and compare the classification accuracies accrued via the application of these two approaches to narrow-band, VHR UAS images collected over two river reaches in Austria. We assess the effect of pixel size on classification accuracy and evaluate performance across multiple dates. Our findings show that the GEOBIA classification accuracies are markedly higher than those of the endmember-based approach, where the former generally have overall accuracies of >85%. Poor endmember-based classification accuracies are likely due to the very small pixel sizes, as well as a large number of classes, and the relatively small number of bands employed.
Geospatial object-based image analysis (GEOBIA) and Endmember-based Land-Cover Classification of Narrow-Band Remote-Sensor Image Data
Category
Poster Abstract
