UAV-Hyperspectral Image Based Mapping in Salt Marshes: A Case Study in the Gomso Bay, Korea

Abstract

The salt marshes are one of the most productively ecological wetlands with high biological productivity and blue carbon sequestration levels. Recently, with the development of image analysis technology for unmanned aerial vehicles (UAV), mapping of salt marsh has become possible more accurately and efficiently. However, the spatial patterns of vegetation species and its aboveground biomass (AGB) in a coastal salt marsh remain unclear. In this study, the RGB and hyperspectral images and LiDAR data were acquired in Gomso Bay from 24 June to 25 June 2021. Two species of Phragmites communis (reed) and Suaeda maritima were widely distributed, and P. communis and S. maritima were dominant in the east and west part of the study area, respectively. DJI Matrice 300 RTK with the Zenmuse P1 RGB sensor (35 mm fixed-focus lens) and L1 sensor, and DJI Matrice 600 RTK with Nano-Hyperspec (Headwall) were used. Hyperspectral image was used to determine the purest spectral signature of each species and LiDAR data was used to determine the canopy of vegetation. Using these methods, it is possible to discriminate between P. communis and reddish S. maritima species with adequate precision. However, greenish S. maritima species did not differ significantly from the spectral signature of P. communis. These two species showed distinct difference in height, two species were classified using LiDAR-based vegetation height. The UAV-hyperspectral and LiDAR data are expected to improve the efficiency of salt marsh mapping. Furthermore, it is expected to significant contribute to estimating accurate blue carbon, away from the method of using only the distribution area information of vegetation in estimating the blue carbon of salt marshes.