ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume IV-2/W5
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., IV-2/W5, 127–132, 2019
https://doi.org/10.5194/isprs-annals-IV-2-W5-127-2019
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., IV-2/W5, 127–132, 2019
https://doi.org/10.5194/isprs-annals-IV-2-W5-127-2019

  29 May 2019

29 May 2019

CAN UAV LIDAR DERIVE VERTICAL STRUCTURE OF HERBACEOUS VEGETATION ON RIVERDIKE?

N. Miura1, T. F. Koyanagi2, S. Yokota3, and S. Yamada4 N. Miura et al.
  • 1Graduate School of Agricultural and Life Sciences, the University of Tokyo, Japan
  • 2Field Studies Institute for Environmental Education, Tokyo Gakugei University, Japan
  • 3Faculty of Environmental Studies, Tokyo City University, Japan
  • 4Faculty of Agriculture, Tokyo University of Agriculture, Japan

Keywords: UAV, LiDAR, Herbaceous vegetation, Grass, Riverdike

Abstract. Herbaceous vegetation on riverdike is important to prevent soil erosion, which may lead to collapse of riverdikes and consequently severe flooding. To keep suitable vegetation condition and secure visibility of riverdike for inspection, managers need to know where thick and tall herbaceous vegetation grows on riverdikes in a cost effective manner. This paper aims to derive vertical structure of herbaceous vegetation on riverdike using UAV LiDAR. UAV LiDAR based indices, V-bottom, V-middle and V-top; presence of vegetation in bottom vegetation (0–50 cm), middle vegetation (50–100 cm) and top vegetation (>100 cm) respectively, were calculated and compared to field survey data for validation. The results showed that UAV LiDAR based assessment strongly correlated with field based assessment in middle and top vegetation with R2 = 0.67 and 0.85, respectively. However, it displayed only moderate correlation in bottom vegetation (R2 = 0.36). It was found that bottom vegetation might be underestimated when forb species or falling grass species is abundant in middle vegetation. However, underestimated place can be easily checked using proposed UAV LiDAR based V-bottom and V-middle values. The vertical structure information derived by proposed method would be valuable assets for vegetation management on riverdike.