A research team led by Kim Won Ok, a researcher at the Faculty of Information Science and Technology, has developed a system for detailed survey of forest resources whereby detailed indices such as the number, height, species and volume of individual trees can be obtained from UAV RGB images.
Past surveys of forest resources involved many investigators on the spot, but the system makes surveys easier, scientific and rapid only by UAV.
The system consists of two parts ― one is for obtaining orthomosaic images, digital surface model (DSM) and digital elevation models (DEM) from UAV images and the other is for detecting individual treetops and classifying tree species based on these data in a certain area.
To guarantee the quality of UAV images for survey, it is important to properly determine the photographing parameters of UAV, so through lots of photographing and analysis of UAV images she set photographing interval as 2~2.5s, photographing altitude as 70m from the hillside of the area and air speed as 2.5~3m/s.
This system ensures over 85% of treetop detection rate and over 85% of classification accuracy, and reduces labor for survey down to one sixth.
© 2021 KumChaek University of Technology