Automatic Plot-wise Tree Location Mapping Using Single-scan Terrestrial Laser Scanning

The application of terrestrial laser scanning (TLS) has received increasing attention in the quantitative forest inventories. Both single-scan and multi-scan TLS can be employed for the forest parameter retrieval. The multi-scan mode captures an ideal data set, which in general provides whole tree coverage and leads to accurate tree trunk detection and modelling. The single-scan data is, however, of high practical interest because of its high collection efficiency and low processing expense achieved by avoiding the multi-scan point cloud matching. In this paper, an automatic plot-wise tree location mapping algorithm by using single-scan TLS is presented. The method detects the tree trunk based on range data and tree trunk model. The test area was a single-layer, pine-dominated forest. The reference data were interactively measured on-screen on the TLS intensity image by an operator. Eighty-five percent of trees that were manually found on the intensity image inside 60m range could be automatically found. The complexity of the forest scene and the occlusion effects are the main limiting factors in the application of the method. The experimental result shows that the single-scan TLS is a feasible technique for plot-wise tree location mapping.