EN
Airborne laser scanning (ALS) technology allows collecting data describing top forest layer in a very accurate way. This provides a possibility to generate crown height models (CHM) with resolution in which single tree crowns can be detected. In presented study highly automatic algorithm for single tree detection is presented. FALCON II ALS acquired data in Forest Experimental Station in Rogów (central Poland). Data used for algorithm accuracy evaluation were acquired by measuring sample plots on VSD photogrammetric station. CHM with the resolution of 0.5 m has been used in the presented research. Algorithm, in automatic or semi− −automatic way, defines single crowns. Each of the end segments was additionally processed for correctly defined size and area of the crown projection. Received accuracy of correctly detected trees was 69% for all samples (71% for pine stands and 61% mixed stands) in automatic method and 74% (78% and 65% respectively) in semi−automatic method.