Light Detection and Ranging (LiDAR) is a laser-based remote sensing technique that measures the two-way travel time of an emitted laser pulse to calculate the distance between the sensor and the observed object. LiDAR systems are frequently mounted on airplanes to gather data of large geographical areas. In this way, LiDAR systems generate tree-dimensional, georeferenced clouds of points of the Earth’s surface. Compared to traditional photogrammetric approach, LiDAR is able to collect data under the vegetation, because more than one reflection from a laser beam can be recognized. Overall, such systems are nowadays capable to perform over 200.000 measurements per second and can obtain over 30 points per square meter. Because of this, LiDAR had become one of the prime technologies for forest surveillance [Koe08]. Obviously, LiDAR datasets are huge and therefore difficult for efficient visualization, archiving, and sending through the internet [Mon10]. Beside this, LiDAR cannot distinguish among different objects on the Earth’s surface. The desired features have to be therefore extracted from the LiDAR datasets by a so-called segmentation process. In practice, this process is still manual or semi-automated and therefore, considerably time-consuming and error-prone.  The pioneering automatic segmentation methods for LiDAR datasets based on points cloud conversion into height maps. In such maps, objects’ contours are firstly searched and then used to facilitate the local planes approximation, and a region growing algorithms to determine categories of considered points [Pri00]. Recent approaches, however, search for topological patterns in point sets and utilize these patterns to classify the points. The final solution is obtained by threshold filtering, neural networks or vector machines [Koe08]. However, the reliability of the actual methods does not exceed 90% leaving enough space for the further research and improvements. The purpose of this project is to join the experience of Austrian and Slovene research groups. The Slovene group has expertise in efficient LiDAR data visualization and compression, while the University of Salzburg has been developing methods for forest analysis, mapping, and management (see [TIE05], [BLA04]). Within the project the joint expertise on the algorithm side and on the application side will be utilized to develop a prototype solution for rendering large size LiDAR data in real-time will be developed by reduction of graphics accelerators workload. This will be achieved by a controlled down-sampling process of faraway points using a quad-tree and the characteristics of the graphical pipeline. A tool for forest data analysis will be developed and included in the visualization system. A suitable compression technique for forestry terrain will be developed, too. It is also envisaged to test transferability of this methodology to calculating solar power energy potentials on the fly. This means that rather than pre-rendering large amounts of point data for solar irradiation maps, intelligent approaches will be developed. These approaches will determine various shadowing effects of existing and planned buildings, trees or other elements on demand out of billions of 3D-points stored in a database. The latter is cutting edge research and will include the Research Studio iSPACE in Salzburg which is an Industry-related research entity under the auspices of the Austrian Ministry for Science.