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

Anthony M Filippi,Andong Ma,Karl-Heinrich Anders,Gernot Paulus,Cesar R Castillo,İnci Güneralp

doi:10.3390/land11020246

Anthony M Filippi, Andong Ma + Show 4 more

Open Access

https://doi.org/10.3390/land11020246

Copy DOI

Abstract

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 for riparian forest mapping. However, for improved riparian forest/zone monitoring, restoration, and management, an enhanced understanding of the accuracy of different classification methods for mapping riparian forests and other land covers at high thematic resolution is necessary. Research that compares classification efficacies of endmember- and object-based methods applied to VHR (e.g., UAS) images is limited. Using the Sequential Maximum Angle Convex Cone (SMACC) endmember extraction algorithm (EEA) jointly with the Spectral Angle Mapper (SAM) classifier, and a separate multiresolution segmentation/object-based classification method, we map riparian forests/land covers and compare the classification accuracies accrued via the application of these two approaches to narrow-band, VHR UAS orthoimages collected over two river reaches/riparian areas in Austria. We assess the effect of pixel size on classification accuracy, with 7 and 20 cm pixels, and evaluate performance across multiple dates. Our findings show that the object-based 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 the large number of classes, and the relatively small number of bands used. Object-based classification in this context provides for effective riparian forest/zone monitoring and management.

Highlights

IntroductionCompared with other terrestrial ecosystems, floodplain forests disproportionately affect the global carbon cycle, since floodplain/riparian forests [1] are an important carbon sink relative to other terrestrial ecosystems [2,3]
This article is an open access articleCompared with other terrestrial ecosystems, floodplain forests disproportionately affect the global carbon cycle, since floodplain/riparian forests [1] are an important carbon sink relative to other terrestrial ecosystems [2,3]
With regard to the poor classification accuracies from the endmember-based approach, this is likely largely attributable to the very small pixel sizes, as well as to the relatively large number of classes used for these classification systems, in conjunction with the relatively small number of bands available with the unmanned aircraft systems (UAS)-derived orthoimages, even though their bandwidths are narrow

Summary

Introduction

Compared with other terrestrial ecosystems, floodplain forests disproportionately affect the global carbon cycle, since floodplain/riparian forests [1] are an important carbon sink relative to other terrestrial ecosystems [2,3]. Such forests can store large amounts of carbon due to high productivity rates and/or saturated conditions that foster belowground carbon storage. Habitats for a myriad of plants and animal species [4] They markedly affect downstream river water quality by minimizing pollution from the surrounding landscape, by enabling increased reduction of nutrients and sediment in higher-biomass areas [5] and by protecting against erosion [6,7,8,9]. For an improved riparian/floodplain forest management, a better understanding of the efficacy of different classification methods for mapping riparian/floodplain forests and other land covers in such areas at high thematic resolution/specificity is needed [9,14]

Objectives

Methods

Conclusion