SEARCH

SEARCH BY CITATION

Keywords:

  • Bog vegetation;
  • Colour infrared;
  • Eriophorum ;
  • Near-infrared;
  • Object-based image classification;
  • Sphagnum ;
  • UAV

Abstract

Question

Can UAV-based NIR remote sensing support restoration monitoring of cut-over bogs by providing valid information on species distribution and surface structure?

Location

Restored polders of the Uchter Moor, a bog complex in NW Germany.

Methods

We used autonomously flying quadrocopters, supplied with either a panchromatic or colour infrared calibrated small frame digital camera to generate high resolution images of the restored bog surface. We performed a two-step classification process of automatic image segmentation and object-based classification to distinguish between four pre-defined classes (waterlogged bare peat, Sphagnum spp., Eriophorum vaginatum and Betula pubescens. An independent validation procedure was performed to evaluate the accuracy of the classification.

Results

A set-up composed of decision rules for reflectance, geometry and textural features was applied for identification of the four classes. The presented classification revealed an overall accuracy level of 91%. Most reliable attribution was obtained for waterlogged bare peat and Sphagnum-covered surfaces, revealing producer accuracies of 95% and 91%, respectively. Lower but still feasible accuracy levels were obtained for Eriophorum vaginatum and Betula pubescens individuals (89% and 84%, respectively).

Conclusions

UAV-based NIR remote sensing is a promising tool for monitoring the restoration of cut-over bogs and has the potential to significantly reduce laborious field surveys. UAVs may increasingly play a significant role in future ecological monitoring studies, since they are small in size, highly flexible, easy to handle, non-emissive and available at a comparatively low cost.