University of Twente Student Theses


UAV RGB images to assess the seasonal effect of canopy on accuracy of DTM and Forest AGB/carbon estimation in Haagse Bos Netherlands.

Worku, Woinshet (2020) UAV RGB images to assess the seasonal effect of canopy on accuracy of DTM and Forest AGB/carbon estimation in Haagse Bos Netherlands.

[img] PDF
Abstract:Light Detection and Ranging technology (Lidar) is a survey tool used for several applications in the field of forestry and forest research. It used to capture the 3D structure of topography and vegetation accurately and quickly cover large areas. Lidar can model the vertical distribution of the canopy and ground surface, which will give full information about vegetation structure, and it estimates the tree crown size, tree height, basal area and stem volume accurately. However, the data set obtained from airborne Lidar is costly to use for regular monitoring and not always accessible. Recently the emerging technology of Unmanned Aerial Vehicle (UAV) is becoming operational in various purposes and applications. This platform is operated from the ground and provide a promising way for timely and cost-effective monitoring of environmental phenomena and natural resources at a very high spatial and temporal resolutions. The major objective of this paper was to assess the seasonal effect of the canopy on the derived DTM, CHM and the outcome of AGB. Thus UAV photogrammetric images captured during the leaf-on and leaf-off season were analysed and assessed how they influence the result. The accuracy of results was evaluated taking Lidar as reference. In this study, the accuracy of DTM generated from UAV RGB images of leaf-on and leaf-off seasons was assessed compared to Lidar derived DTM. The result obtained indicates that the accuracy of DTM from UAV images of the leaf-off season was comparable with the Lidar derived DTM and showed a strong correlation. The RMSE error of leaf-off season was 0.25. But, the accuracy of DTM during the leaf-on season was declined, and an error increased to 0.3. The UAV leaf-on and leaf-off seasons DTM has 80% and 72% correlation with reference DTM, respectively. The comparison of tree heights extracted from UAV RGB images of leaf-on and leaf-off season has shown a strong correlation with RMSE and R2 of 2.2 m and 0.88, which show that tree height measurements explain 88 % of the difference in height measurement in UAV leaf-on season from leaf-off season data. Similarly, the AGB/AGC computed from UAV leaf-on and leaf-off season has good agreement with each other and resulted R2 of 0.96 with RMSE of 0.13 and 0.06 Mg/tree respectively. The accuracy assessment of AGB derived from Lidar and UAV datasets has shown a positive and strong correlation. The comparison of Lidar AGB with UAV leaf-on season has shown R2 of 0.85 with RMSE of 0.724 Mg/tree. Whereas, UAV leaf-off season has shown R2 and RMSE of 0.91 and 0.72 Mg/tree respectively. Keywords: leaf-on; leaf-off; Height; UAV; AGB; DTM; CHM; Lidar
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page