Unmanned Aircraft System Photogrammetry for Mapping Diverse Vegetation Species in a Heterogeneous Coastal Wetland
- PDF / 2,272,342 Bytes
- 13 Pages / 595.276 x 790.866 pts Page_size
- 96 Downloads / 258 Views
WETLANDS RESTORATION
Unmanned Aircraft System Photogrammetry for Mapping Diverse Vegetation Species in a Heterogeneous Coastal Wetland Sara Denka Durgan 1 & Caiyun Zhang 1
&
Aaron Duecaster 1 & Francesca Fourney 2 & Hongbo Su 3
Received: 12 March 2020 / Accepted: 1 September 2020 # Society of Wetland Scientists 2020
Abstract Acquiring detailed information on wetland plant species is critical for monitoring wetland ecosystem restoration and management. The emerging technique of Unmanned Aircraft System (UAS) photogrammetry has immense potential for such applications. In this study, we assessed the capacity of UAS photogrammetric products for classifying and mapping a large number of wetland plant species using contemporary Object-Based Image Analysis (OBIA) and machine learning methods. Our testing results in a heterogeneous coastal wetland demonstrated the benefit of centimeter-level orthoimagery and vertical products from UAS photogrammetry for mapping 17 species compared with standard aerial photography products. We achieved an overall accuracy (OA) of 71.3% and 84.8% for mapping 17 species and 10 major species, respectively. Our study suggests that UAS photogrammetry is a valuable tool for mapping wetland species composition and distribution. Keywords UAS . Coastal wetland . Structure from motion . Species mapping . Object-based classification
Introduction Coastal wetlands are considered one of the most productive ecosystems; yet are under increasing threat by anthropogenic and climate stressors such as fragmentation, diversion of freshwater input, and sea level rise (Browder et al. 1994; Kirwan and Megonigal 2013). Mitigating the loss of these vital resources requires the development of efficient monitoring procedures to focus restoration efforts. Species composition is an important metric for determining the ecological integrity of a wetland ecosystem and informing management strategies (Ramsey and Jensen 1996; Lee et al. 2006; Zhao et al. 2016). Vegetation species maps not only serve as a proxy for the overall health of the ecosystem, but they also reveal information on soil conditions and water quality (Morris et al.
* Caiyun Zhang [email protected] 1
Department of Geosciences, Florida Atlantic University, 777 Glades Road, Boca Raton, FL 33431, USA
2
Cyriacks Environmental Consulting Services, 3001 SW 15th Street, Suite B, Deerfield Beach, FL 33442, USA
3
Department of Civil, Environmental and Geomatics Engineering, Florida Atlantic University, Boca Raton, FL 33431, USA
2002). The ability to generate this data in a timely and costeffective manner is critical for managers to track the progress of restoration activities. Remote sensing has been utilized for mapping wetland vegetation over the past several decades. Traditional remote sensing platforms (i.e. satellites and manned aircraft) are ideal for global and regional-scale mapping, but local-scale, fragmented, and diverse wetlands require higher spatial resolution images to capture small vegetation units (Adam et al. 2010). Unmanned Aircraft Sys
Data Loading...
