Estimating the quality of roundwood and tree trunks using LiDAR

PI : Alexandre Piboule (ONF – Pôle R&D Nancy)

Co-applicants : Laboratoire d’Etude des Ressources Forêt-Bois (LERFOB)

Collaboration :  Thièry Constant, Francis Colin


Context — The ground-based LiDAR technology (Light Detection And Ranging) allows the description of complex forestry scenes, with a level of detail, today, unequalled. By scanning its environment by a laser beam until one hundred meters, the device measures the distance to the closest obstacle in millions of directions with an accuracy of some millimetres. Important issues are based on this technique which shakes up the scientists’ or managers’ ability for measuring trees at plot scale. Numerous recent scientific publications demonstrate the general interest for the use of this technique in forestry.

The ONF (French Forestry Department) through a collaboration with ENSAM (Cluny) initiated the development of a software platform called Computree (http://computree.onf.fr/) which aims to pool the computing developments linked to the use of the LiDAR technology, and to facilitate its usage in forestry.

Objectives — The project objective is to extract from the information delivered by a ground-based LiDAR, some characteristics of the global shape of the trunk (volume, curvature, inclination…), and to complete them by characteristics of defects at the trunk surface which modify its relief. These defects range from the presence of a branch, to knots, and until a slight relief variation corresponding to a branch scar in some favourable conditions of acquisition.

Approach — Our proposal is to develop algorithms allowing (i) the detection of suspected zones by an analysis of the local relief, then for each zone (ii) to identify the defect type, and (iii) to determine its characteristics in order to integrate them into grading rules. These algorithms will be developed within the Computree framework. The quality of the algorithms will be assessed from existing data: for several hundreds of logs, these data combine LiDAR information describing the bark surface, and the internal nodosity provided by X-Ray scanning.

Expected results and impacts — The development of an automated method providing quality criteria for grading standing trees would bring an undeniable added value for the monitoring of a forest resource, whether for inventory, for the ecosystem functioning or for commercial purposes.