Uporaba lidarskih podatkov za klasifikacijo pokrovnosti

Abstract

The book presents a method to create a land cover map from lidar data (elevation, intensity and standard deviation of height). The intensity of the observed object depends on many factors and is therefore not easy to interpret. With careful study of the problem and implementing various improvements we managed to distinctively classify categories of grass, agricultural fields and asphalt. We used standard deviation of height to differentiate trees and buildings, because this was not possible using intensity data alone. On the basis of independent lidar data a high quality three dimensional land cover map of local area has been successfully generated. It distinguishes five basic categories, although more detailed sub-categories could be introduced if necessary. Once the methodology is determined, products can be generated fully automatically, with minimal effort and costs. The resulting maps are of very high positional and thematic accuracy with numerous advantages for local studies.