Identification of Scandinavian Commercial Species of Individual Trees from Airborne Laser Scanning Data Using Alpha Shape Metrics

• DOI: 10.1093/forestscience/55.1.37
• Publication Date: 2024-04-10T09:28:40Z
• Authors: Jari Vauhkonen; Timo Tokola; Petteri Packalén; Matti Maltamo
• Keywords: Laser Scanning; Identification; Alpha (finance); 3D Scanning

Abstract Airborne laser scanning (ALS) data are not usually considered to be very informative with respect tree species, and this information is often obtained by combining such spectral image material. The aim here was test the ability of variables derived solely from ALS describe crown shape structure characteristics required for species discrimination. For that purpose, we constructed approximations three-dimensional return applying a computational geometry approach, alpha concept, developed metrics describing them. We examined these classify Scandinavian commercial (pine, spruce, deciduous trees) means linear discriminant analysis compared groups ALS-based height, density, intensity, two-dimensional texture variables. evaluating classification accuracy, used set composed 92 dominant or codominant trees detected delineated manually density approximately 40 returns m−2. proved capable discriminating between all three classes evaluated, several height distribution were found discriminate coniferous species. An overall accuracy 95% κ coefficient 0.90 achieved using combination Because initial application carried out on only clearly detectable remotely sensed data, further research apply approach presented within stands continuous canopy. Furthermore, as observations highly biased toward mature trees, experiments more representative sets needed generalize result obtained.