Non‐destructive estimation of above‐ground surface and near‐surface biomass using 3D terrestrial remote sensing techniques

Abstract

Summary Quantitative measurements of above‐ground vegetation biomass are vital to a range of ecological and natural resource management applications. Remote‐sensing techniques, such as terrestrial laser scanning (TLS) and image‐based point clouds, are potentially revolutionary techniques for measuring vegetation biomass and deriving other related, structural metrics for these purposes. Surface vegetation biomass (up to 25 cm) in pasture, forest, and woodland environments is estimated from a 3D point cloud derived from a small number of digital images. Volume is calculated, using the 3D cloud and regressed against dry weight to provide an estimate of biomass. Assessment of the method is made through comparison to 3D point clouds collected through TLS surveys. High correlation between destructively sampled biomass and vegetation volume derived from TLS and image‐based point clouds in the pasture (TLS r2=0·75, image based r2=0·78), dry grassy forest (TLS r2=0·73, image based r2=0·87) and lowland forest (TLS r2=0·74, image based r2=0·63) environments was found. Occlusion caused by standing vegetation in the woodland environment resulted in moderate correlation between TLS derived volume and biomass (r2=0·50). The effects of surrounding vegetation on the image‐based technique resulted in 3D point clouds being resolved for only 40% of the samples in this environment. The results of this study demonstrate that image‐based point cloud techniques are highly viable for the measurement of surface biomass. In contrast to TLS, volume and biomass data can be captured using low‐cost equipment and relatively little expertise.

Extraction of Sample Plot Parameters from 3D Point Cloud Reconstruction Based on Combined RTK and CCD Continuous Photography

“…Remote sensing images and aerial photogrammetry promptly obtain forest surface data over a large area. However, these two methods have a limitation because the accuracy of the terrestrial data under the tree crown is compromised [10][11][12][13]. 3D laser scanning technology can efficiently collect high precision 3D data of a forest's vertical structure.…”

Investigating Surface and Near-Surface Bushfire Fuel Attributes: A Comparison between Visual Assessments and Image-Based Point Clouds

“…Surface fuel load was calculated for both visually assessed and point cloud data by inputting estimates of surface litter depth (height) and percent cover obtained via the two methods into the following Equation (2) as per Wallace et al [40]:$Fuel load =\frac{1.46 \times height \times cover}{\left(\pi \times {10}^2\right)+0.98}$…”

“…However, technologies such as LiDAR and TLS can be cost prohibitive [38], and are therefore limited in their application. Recent studies have demonstrated the value of more cost-effective methods of capturing information describing vegetation (fuel) structure and health using image-based point clouds [39,40,41,42,43,44,45]. Image based point clouds are constructed using photogrammetric principles from highly overlapping photography, and the reader is directed to Westoby et al [46] and Snavely et al [47] for further details on the methods.…”

Surveying pasture communities in diachronic analyses by 3D models: the diachronic canopy variation model

“…Recent comparisons of AGB estimations for pasture communities resulted from SfM and TLS techniques (Wallace et al. , Cooper et al. ) showed that both methods allow performing an accurate estimation of the surface biomass.…”