An enhanced algorithm for the detection of individual trees based on airborne laser scanning data combining geometric and dendrometric rules

Airborne laser scanning has become a well-established, consolidated, and widely used technology for individual tree detection. In this context, we have developed the geoprocessing tool xTree to allow for the detection of individual tree tops and tree crowns based on a combination of airborne laser scanning data and a priori information consisting of geometric and dendrometric rules. The primary aim of this article is to present the workflow for the detection of individual trees using the xTree tool. A detailed description, flowchart, and scheme are provided for these purposes. The additional aim is to present the performance of this workflow through a case study. The study area covered 2.8 ha of forests in Slovakia. The airborne laser scanning data reached an average density of 18 ±6 points per m². The ground reference data included 2,119 trees in coniferous, broadleaved, and mixed strata. The enhanced detection algorithm achieved overall correctness of 85 ±2%. The overall completeness from the tree number was 69 ±4%, and the overall completeness from the tree volume was 78 ±2%. Generally, the performance increased with the increasing proportion of coniferous trees and diameter of trees.