Epsilon Archive for Student Projects

Abstract

Laser scanning of forests has recently become a bit of a revolutionary tool in forestry. It is a smooth and cost-effective tool while it is usable over very large areas. The information extracted from the laser scanning provides a good overview of the property but is also a good basis for decision making of future forestry measures and how these are to be carried out.
The purpose of the study was to find out the surface model measurement accuracy at stock level. To be able to use surface model as a basis for decision making of forest measures, it is necessary to ensure the material's measurement accuracy, current overestimation / underestimation at the stock level.
The study was conducted on thinned spruce dominant stocks in Skåne with an age range of 20-80 years. The sample surfaces were randomized into the stocks whose size was at least two hectares. Within the test areas, all strains were clawed and the height was measured on every other tree within each two centimeter class. The results from the field study were then compared to the results of the laser scanning, where it could be seen that the laser scanning height model underestimates the height of the stock with an average height of 3,15 meters (16%). The underestimation cannot be tied to the population's number or height, however, the surface model shows a greater underestimation in the areas containing large openings. One theory to this result is that more laser points take directly on the ground and thus hit high undergrowth vegetation, protruding branches etc. in this areas whit more openings, which in turn lowers the surface models average height.