The Influence of Forest Cover Characteristics on the Accuracy of LiDAR Based Digital Terrain Model

Abstract

The influence of forest cover characteristics on the accuracy of LiDAR based altitude of digital terrain model is analyzed on a test area in central part of Lithuania. Two variants of LiDAR point clouds, based on 1 and 6 poiiils/m1 scanning densities were used lo extract the digital terrain model. Circular sample plots were located and measured in the forest using conventional geodetic survey and dendrometric methods. Higher laser scanning point density was found to result in significantly lower bias of altitude of digital terrain model lhan the lower laser scanning point density. The bias of the altitude of digital terrain model was less on the sample plots located in coniferous forest than in mixed or deciduous forests. Presence of the 2ni storey and undergrowth was not found to have significant influence on thir bias of altitude of digital terrain model, no matter the laser scanning characteristics. A constrained linear ordination meiliod (redundancy analysis) indicated that the biases of altitude of digital terrain model were negatively related with the magnitude of major volumetric characteristics of the I "storey of ihejjtand and share ol coniferous trees but positively with the one of the 2"d storey