OPS 4-17
Seeing forests from drones: Testing the potentials of drones in a subtropical forest

Tuesday, August 11, 2015
Exhibit Hall, Baltimore Convention Center
Jianbo Hu, Ministry of Transport, Tianjin Research Institute of Water Transport Engineering, Tianjin, China
Jian Zhang, Renewable Resources, University of Alberta, Edmonton, AB, Canada

      Unmanned Aerial Vehicle (UAV) commonly known as a drone, is becoming more and more common, used as a tool to overlook the earth in the sky. UAV could provide much higher resolution images than satellite, and also much easier, cost-effective and more portable than manned aerial vehicle.

      Like other UAV application areas, forest ecology could also benefits from UAV to investigate the feature of forest canopy, like tree species distribution, canopy size measurement, and etc. It will be very time-consuming and even infeasible for ecologist to do such a work in a dense forest. The questions are what kinds of specific tasks could UAV do, which kind of UAV is suitable, and how to fly and perform the task in a safe way.

      In 2014, we tested a electrically powered quadrocopter UAV (MD4-1000) in the Dinghushan National Nature Reserve (subtropical forest) in southern China. We did photogrammetric photography and panoramic photography to cover the 20-ha Dinghushan subtropical plot. In this paper, we will detail the test and share the experience.


      The restriction and risk of UAV application in forest ecology is limited takeoff and landing site. Rotor UAV is much better than fixed-wing UAV if the area is no more than 1 km2. Our good fortune is a temple next to the plot, so that we set the takeoff and landing site in the open space of the temple. If not, we have to cut trees and clean out enough space (about 100 m2).

      Photogrammetric photography could provide high resolution orthoimage and Digital Surface Model (DSM). We can distinguish and locate dead trees, different kinds of tree species, and measure the size of tree canopy and forest gap in the high resolution orthoimage. Theoretically, we could also obtain tree height parameter by subtracting the DEM acquired from field survey, however it depends on the DEM quality, since it was acquired yeas ago and not targeted for UAV application.

     Panoramic photography could provide a panorama photo covering the whole plot. Panoramic photography is much more easier than photogrammetric photography (professionals and professional software), however panorama photo is only for watch not for measurement. The reason to watch using panorama photo instead of orthoimage is that orthoimage has problems at stitching lines.