• Home
  • Media
  • News
  • Australian and New Zealand forest researchers collaborate to extract information from LiDAR dense point clouds

Australian and New Zealand forest researchers collaborate to extract information from LiDAR dense point clouds

Dr Christine Stone – Leader NSW Forest Science (Dept of Industry –Lands) and Dr Michael Watt – Research Leader Geometrics  (SCION)
 
Scientists from several Australian research agencies and NZ’s SCION (formerly Forest Research Institute) are collaborating in a multidisciplinary FWPA project titled “Optimizing remotely acquired, dense point cloud data for plantation inventory”.
 
The Riegl VUX-1 Lidar scanner which was mounted onto a helicopter and flown over plots within the Snowy Region of the Forestry Corporation NSW.
This collaborative Australian/New Zealand project brings together experts across a range of disciplines that do not reside collectively in any one institution. The project team has internationally recognised expertise in LiDAR & UAV technologies, software engineering and sophisticated modelling approaches. Participating research groups include the Forest Informatics team at SCION, the Terra Luma and photogrammetry research groups at the University of Tasmania, the Australian Centre forField Robotics at the University of Sydney and the  NSW Lands Forest Science team, as well as two high profile forest service providers, Interpine and Indufor Asia Pacific.
 
This collaborative Australian/New Zealand project brings together experts across a range of disciplines that do not reside collectively in any one institution.
 
Prototype UAV platform configured by the Terra Luma group (University of Tasmania) for carrying the Velodyne LiDAR sensor.
Significant progress has been made in the operational adoption of remotely acquired data, in particular LiDAR data, by plantation growers both in Australia and in New Zealand for the assessment of plantations. Two previouslyfunded FWPA projects have helped drive this paradigm shift in company awareness and their decision to integrate this new technology into their planning management systems. The modelling and data workflow processes developed in these projects utilise data acquired by established, commercial LiDAR sensors. The rapid advances in this technology present numerous, potential applications for improved cost efficiencies and assessment precision for timber plantation growers. New sensors can now generate 3D datasets with densities greater than 100 points/m2 including the Riegl VUX-1 LiDAR scanner (with 10mm survey-grade accuracy) and the Velodyne Lidar for UAV platforms (Figures 1a and b, 2 and 3). Datasets acquired by both these new sensors have been acquired for the project and future UAS acquisition campaigns are scheduled for both Tasmanian and NZ study sites.
Prototype UAV platform configured by the Terra Luma group (University of Tasmania) for carrying the Velodyne LiDAR sensor.

The overall aim of this project is to develop and evaluate robust, optimal workflow solutions for the processing and analysis of dense point cloud datasets acquired from both LiDAR airborne and multi-rotor UAV systems flown over Pinus radiata plantations. An important component of this process is evaluation of the geometric precision of these dense point cloud datasets in order to determine if the point location accuracies are comparable with estimates derived from the manually measured trees.
The researchers will compare a suite of 3D visualisation software packages for their potential for on-screen individual tree assessment. As part of this process, collaboration with the University of Tasmania’s virtual reality laboratory is being considered. Another objective is to develop reconstructed 3D models of individual trees to estimate log product outturn. Programming techniques developed within the robotics sector will be explored as part of this objective. Finally by December 2017 the project intends to deliver recommendations to the plantation sector that will identify cost-effective options for the practical utilisation of this technology.