Remote sensing and Terrestrial LiDAR: Shaping the future of forestry

A remote sensing workshop, with a specific focus on terrestrial lidar phenotyping, was held on 15 May 2024 at the Karkloof Club outside Howick, KZN. The workshop, organised by Camcore and Sappi, included participation from local Camcore member companies Mondi, MTO, PG Bison, Sappi and York Timbers, as well as other organisations involved in relevant remote sensing projects. Camcore, an international tree breeding and gene conservation co-operative, established in 1980, based at North Carolina State University in Raleigh North Carolina USA.  The Co-operative includes memberships from South America, Africa and Indonesia. Camcore’s focus in the early years was on the testing and gene conservation of various indigenous Pine species occurring in Central America and Mexico.  They have since broadened their focus to include Eucalypts and other hardwoods and are also actively involved in the development of enabling technologies with its members. One of the current focus areas is the use of remote sensing and new imaging technologies to assess important characteristics in genetic trials. The purpose of the workshop was to provide a review of current remote sensing projects and form a LiDAR working group among practitioners. Presentations were delivered by staff from Camcore, University of Pretoria, University of Stellenbosch, ICFR, Sappi, Mondi and York Timbers, and a hands-on coding course using RStudio was also presented during the afternoon.

Terrestrial LiDAR

Forestry management is entering a new era of precision and efficiency with the advent of advanced technologies. One such technology is terrestrial Light Detection and Ranging (LiDAR) which is transforming how we measure and understand forests. This technology uses laser pulses to capture detailed three-dimensional point clouds, that can be reconstructed into trees and forest stands, allowing for the extraction of multiple features from a single scan. In addition to the traditional dendrometric measures of diameter at breast height (DBH) and tree height, complex metrics that were previously difficult to measure can be captured. Terrestrial LiDAR provides a platform for simultaneously capturing numerous phenotypic metrics in a single scanning campaign, including:

• Features such as leaf area index and branching angles that are crucial for understanding tree architecture, canopy structure and light interception patterns. This complex structural data can provide further insight into forest ecosystems, particularly how trees compete for sunlight and how this impacts growth, a key factor in forest dynamics.
• Direct measurements of tree volume. Tree volume estimates, derived from allometric relationships of height and DBH, can lead to inaccurate commercial volume estimates but LiDAR can improve accuracy by providing direct measurements of tree volume.
• Surface features of the stem such as branch scars, knots and bark texture, can serve as proxies to gauge trunk quality and distinguish tree species.

Progress and Challenges

To demonstrate the feature extraction process participants were provided with access to LiDAR point cloud data obtained from Pinus taeda progeny trials. This data set was used to demonstrate how metrics such as DBH, tree height, and volume can be extracted from point cloud data using the open-source “FORTLS” and “lidR” packages in R. The packages “ggplot2” and “rgl” were used to construct 3D plots of LiDAR derived DBH and height, overlayed onto positional information per tree family (Figure 2A).

Comparison of the direct field measurements to the LiDAR derived metrics, revealed that DBH is more accurately measured than height. This may arise due to occlusions caused by the tree crown that introduce noise in the foreground, leading to poor differentiation of the upper stem (Figure 2B, C, D). Removing noise from leaves and branches within and between tree canopies are areas that require further development.

Terrestrial LiDAR technology holds immense promise for revolutionizing forest inventory and tree phenotyping, however, scanning campaigns are often case specific, and yet to be optimised. As discussed during the workshop, scanning at higher resolutions or including multiple scan points is often accompanied with the caveat of increased data loading and processing times.

To address these issues, the Camcore R code was designed to extract individual trees using georeferencing information. This significantly reduced the computational power necessary to reconstruct the point clouds. This iterative technique can be applied to extract multiple trees from larger blocks of trees, depending on the specified coordinates (Figure 3).

The number of scan positions as well as acquisition speed of the LiDAR will require consideration to determine the minimal scanning time required to obtain accurate information. If a clear framework for scanning campaigns can be developed, this can be used to improve signal-to-noise ratio, reduce data quantity and computational requirements. This means the user must have clearly defined objectives, while taking into account post-processing capabilities. Ultimately terrestrial LiDAR is an evolving technology in the forestry space, and offers the opportunity to efficiently capture forest inventories, saving time and resources. The detailed data it provides supports better decision-making for thinning, optimizing space for competition, and assessing tree form. As demonstrated in the workshop, this technology provides valuable phenotypic data for genetic studies and tree selection, enhancing breeding programs. LiDAR can also assess tree flowering patterns and seed production, offering insights crucial for both conservation and commercial forestry. Evidently, the features captured by terrestrial LiDAR can inform precise end use of tree resources, while providing intricate phenotypic information for breeding and selection programs.

The Camcore team highlighted that future pipelines will focus on obtaining more complex features including tree form, total biomass, and branch angles. Participation in the workshop emphasized the importance of collaboration to achieve these goals. Shared experiences, data and knowledge transfer are crucial for advancing the use of LiDAR in forestry, making it a more reliable and accessible tool for all.

Source: LiDAR

Back to NEWS:
BUSINESS  l  CLIMATE CHANGE  l  CONSTRUCTION  l  EDUCATION  l  ENVIRONMENT  l  FORESTRY ENGINEERING  l  FORESTRY  l  GOVERNMENT  l  INTERNATIONAL  l  LAND  l  RECYCLING l  RESEARCH  l  ROOFING  l  SHORT HAUL  l  SILVICULTURE  l  SOCIAL RESPONSIBILITY  l  TRANSPORT  l  TREATMENT  l  TRANSPORT  l  VALUE ADDING