Terrestrial laser scanning for forest inventories

20 December 2017

Terrestrial laser scanning (TLS) is an effective technique for acquiring detailed tree attributes in forest plots. During the last two decades, tremendous effort by national mapping agencies, companies, universities and research organisations has been put into developing methods for tree attribute estimation using TLS. There is, however, still a lack of proper understanding on TLS performance.

Different data collection methods and processing standards have led to a large range in tree detection and measurement accuracy. This article explains the early results of an international benchmarking initiative for TLS methods in forest inventories. The study has identified important differences in methods that should lead to operational work guidelines. Measuring Tree Attributes

A terrestrial laser scanner automatically documents its surrounding environment in three-dimensional (3D) space with millions to billions of 3D points. In forestry, TLS is an effective technique for measuring forest plots and is anticipated to be used in national forest inventories, leading to more sustainable silviculture and savings for forest owners and industry alike.

During the last two decades, significant research has been conducted on developing best practices around TLS for forest inventories – to evaluate, for example, whether one scanning position at the plot centre (single scan) or several scanning positions inside and outside of the plot (multi-scans) should be used to measure a sample plot and estimate tree attributes (tree height, diameter, taper, crown width).

Impressive results have been reported in recent years that are automatic, repeatable, accurate for practical applications and comparable to results from national allometric models. There is, however, still a lack of proper understanding on TLS performance, especially in forests with varying structure and development stages (complex forest structures).

Currently, the results obtained from TLS data for plot-wise tree attribute estimation have varied significantly from study to study. The percentage of correctly detected trees from reported multi-scan data has ranged from 50 to 100%. The differences between varying detection rates arise from different TLS hardware, scanning set-up, forest structures and processing methods.

Benchmarking Study

To clarify the current status of the TLS application in plot inventories, an international benchmarking study was launched in 2014, led by EuroSDR and partly funded by the European Community’s Seventh Framework Programme. The main objective of this current benchmarking study is to understand recent developments of TLS methodologies in plot inventories by comparing and evaluating the quality, accuracy and feasibility of the automatic and semi-automatic tree extraction methods based on TLS data.

To read more on the study and progress, click here.