TREE HEIGHT ESTIMATION METHODS FOR TERRESTRIAL LASER SCANNING IN A FOREST RESERVE G. Király a *, G. Brolly a a Univ. of West H, Dept. of Surveying and Remote Sensing, 9400 SOPRON, Bajcsy-Zs. 4, Hungary (kiraly.geza, gbrolly)@emk.nyme.hu Commission III, WG III/3, VIII/11 KEY WORDS: Terrestrial Laser scanning, Forestry, Photogrammetry, Modelling, Single tree ABSTRACT: A terrestrial laser scanning survey has been performed on a forest reserve site. This paper represents the processing of the dataset. Two new methods have been developed for tree positioning, which is necessary for height calculation. One of them is the clustering method, the other is the so-called crescent moon method. The stems were modelled, as high as possible, by both methods. Two methods were tested for height estimation. One was a direct method based upon the Digital Surface Model; however a new method was developed to predict the tree-top based upon the modelled trunk and then the nearest local maximum was selected to that predicted point in the DSM. The other was an indirect method, where the modelled stems were described as paraboloids, and the heights were extrapolated. The accuracies of the methods were investigated by means of photogrammetric measurements based on very high-resolution aerial images. Both the clustering and crescent moon methods were appropriate in tree positioning and stem modelling. The height estimation based upon the DSM yielded good results. The extrapolation method was tested for European Larch only. The height in one-third of the cases was estimated by this method within tolerance. This method can be taken into consideration in the case of leaves on scanning in very dense, multi-storey forests. 1. INTRODUCTION Sixty-eight forest reserves have been established in Hungary to protect and study natural or semi-natural forest ecosystems. Some of these reserves are selected for long-term investigations, and a permanent sampling network has been established. These network points are the geometric base for the collection of different data, such as the forest stand parameters, soil, vegetation and so on (Horváth et al. 2001). Single-tree mapping is not the main objective of these networks, but some of them include stem maps created with conventional instruments, and demand is growing for remote sensing methods producing single tree maps. Terrestrial laser scanning is an available opportunity to automate data acquisition in forests. Several researchers have reported promising results in applicability in forest inventories, e.g. Thies & Spiecker (2004), Aschoff et al. (2004). Forest Reserve 46, situated next to the Austrian border, was selected for this study. This reserve is a mountainous mixed, almost coeval forest stand with oak, hornbeam, beech, larch and spruce species. The spruces are dying of bark beetle disease. 2. MATERIAL 2.1 TLS data The terrestrial laser scanning (TLS) was carried out on a permanent sample plot, No. 05-11, in leaf-off state, in April 2006 with a Riegl Z420i instrument. The configuration of the data acquisition can be seen in Figure 1. The scanning was done on the whole upper hemisphere at three different tilting angles, resulting in a vertical range from -40° to 90° with a 0.055° angle resolution. There were three further surveying points from which the surveys were done toward the sample plot at approx. 180° horizontal, and from -40° to 90° vertical angles with a 0.12° angle resolution. The data acquisition, the registration and the colour coding were performed by the surveying company. Figure 1. The configuration of the TLS data acquisition 2.2 Aerial images Very large scale, colour metric aerial photographs were taken in April 2005. The parameters of the photography can be seen in Table 1. 211 IAPRS Volume XXXVI, Part 3 / W52, 2007