Please note that this work is an extended version of the introduction and literature of the following publication:
Miltiadou, M., Campbell, N. D., Gonzalez Aracil, S., Brown, T. and Grant, M. G. (2018), `Detection of dead standing Eucalyptus camaldulensis without tree delineation for managing biodiversity in native Australian forest', International Journal of Applied Earth Observation and Geoinformation 67, 135-147.
Full Paper Available here: https://www.researchgate.net/publication/323398945_Detection_of_dead_standing_Eucalyptus_camaldulensis_without_tree_delineation_for_managing_biodiversity_in_native_Australian_forest
The importance of Dead Wood
The value of dead trees from a biodiversity management perspective is large. Once a tree dies, its woody structure remains for centuries and it contributes to forest regeneration while providing resources for numerous surrounding organisms (Franklin et al., 1987). More than 4000 species inhabit dead wood in Finland (Siitonen, 2001), where an estimate of 1000 species are threatened (Hanski, 2000). These species include animals, birds and other organisms, like fungi. Fungi contributes to wood decaying, formation of hollows and biodiversity, which supports the resilience of our ecosystem (Peterson et al., 1998). Observing the changes of fungal diversity on decaying wood has an increased interest in science (Abrego and Salcedo, 2011) (Stokland and Larsson, 2011) (Lonsdale et al., 2008) in order to ensure the continuous existence of decaying wood in forests.
In Australia, tree hollows play a signi cant role in managing biodiversity (Lindenmayer et al., 1997)
(Bennett et al., 1994). Nearly all arboreal mammals rely on hollows with the exception of the Koala (Phascolarctos cinereus) and perhaps Ringtail Possums (Pseudocheirus peregrinus) that preferentially make a stick nest. Additionally, numerous Australian bird species use hollows for shelters (Gibbons and Lindenmayer, 2002). Nevertheless, Australia has no real hollow creators unlike the northern hemisphere (e.g. Woodpeckers), and therefore it relies predominantly on natural processes of limb breakage, insect and fungal attack when access points are provided through damage caused by wind, storms and re. This kind of hollows takes hundreds of years to form (Wormington and Lamb, 1999).
According to Gibbons et al. (2000), hollows are more likely to exist on dead trees trees or trees in poor physiological condition. In Australia, studies predict shortage of hollows for colonisation in the near future (Lindenmayer and Wood, 2010) (Goldingay, 2009). A sample list of species that rely on hollows, provided by Forestry Corporation of NSW, is depicted at Figure 1. Three of them are threatened (New South Wales Government, 2016). Consequently, automated detection of dead trees plays a substantial role in managing biodiversity.
 |
Figure 1: Some species that uses tree hollows for shelters. The red ones / bold ones are threatened: Kook-
aburra, Sulphur Crested Cockatoo, Corella, Crimson Rosella, Eastern Rosella, Galah, Rainbow Lorikeet,
Musk Lorikeet, Little Lorikeet , Red-winged Parrot, Superb Parrot, Cockatiel, Australian Ringneck (Par-
rot), Red-rumped Parrot, Powerful Owl, Sooty Ow, Barking Owl, Masked Owl, Barn Owl, White-throated
Treecreeper, Hollow Owl, Brush-tailed Possum |
As explained above, monitoring dead trees is essential for preserving a resilient ecosystem. Remote
sensing automates the process of monitoring forest and increases the spatial resolution of the monitored area.
Related Work in Remote Sensing
Remote sensing was introduced for automating detection of dead trees since fieldwork is a time consuming task, considering the variance spread of trees and the spatial resolution of the area of interest. From a classification perceptive, the task of identifying dead standing and dead fallen trees is di erent. Fallen trees are identi ed by detecting segments or line-like features on the terrain surface using LiDAR (Polewski et al., 2015) (Mcke et al., 2013). Regarding standing dead trees, their shape (reduced number of leaves or broken branches) (Yao et al., 2012) and light reflectance (less green light illuminated) (Pasher and King, 2009) areimportant factors for identifying them.
Previous work on dead standing trees detection performs single tree crown delineation before health assessment (Yao et al., 2012) (Shendryk, Broich, Tulbure, McGrath, Keith and Alexandrov, 2016). Tree crown delineation is usually done by detecting local maxima from the canopy height model (CHM) and then segmenting trees using the watershed algorithm (Popescu et al., 2003). Improvements has been achieved by introducing markers controlled watershed (Jing et al., 2012) and structural elements of tree crowns with di erent sizes (Hu et al., 2014). Additionally, Popescu and Zhao (2008) analyse the vertical distribution of the LiDAR points in conjunction with the local maximum filtering of CHM.
In the case of Eucalyptus in Australia, tree delineation is a challenge due to their irregular structure and multiple trunk splits. Local maxima filtering, used for tree detection, leads to over-segmentation because each tree trunk split forms a local maxima. Shendryk, Broich, Tulbure and Alexandrov (2016) published an interesting Eucalyptus delineation algorithm that performs segmentation from bottom to top; the trunks point cloud is separated from the leaves and individual trunks are identified before the segmentation. Nevertheless, the density resolution starts from 12 points/m2 and goes up to 36 points/m2 around forested areas. For small research projects capturing this high resolution is reasonable, but for larger areas, the density of the emitted pulses is above the optimal resolution for a cost effective versus quality acquisition (Lovell et al., 2005). Miltiadou et al. (2018) presented an new research direction for forest health assessment without tree delineation that uses 3D windows for extracting structural features and using these structural features to train an object detection system. This works was extended in using multi-scale 3D windows for tackling height differences (Miltiadou et al., 2020).
References
Abrego, N. and Salcedo, I. (2011), `How does fungal diversity change based on woody debris type? a case study in northern spain',
Ekologija 57(3). doi: 10.6001/ekologija.v57i3.1916.
Bennett, A., Lumsden, L. and Nicholls, A. (1994), `Tree hollows as a resource for wildlife in remnant woodlands: spatial and temporal patterns across the northern plains of victoria, australia',
Paci c Conservation Biology 1(3), 222-235.
Franklin, J. F., Shugart, H. H. and Harmon, M. E. (1987), `Tree death as an ecological process',
BioScience 17(8), 550-556. doi: 10.2307/1310665.
Gibbons, P. and Lindenmayer, D. (2002), `Tree hollows and wildlife conservation in australia',
CSIRO Publishing . doi: 10.5860/choice.40-1547.
Gibbons, P., Lindenmayer, D., Barry, S. C. and Tanton, M. (2000), `Hollow formation in eucalypts from temperate forests in southeastern australia',
Paci c Conservation Biology 6(3), 218.
Goldingay, R. L. (2009), `Characteristics of tree hollows used by australian birds and bats',
Wildlife Research 36(5), 394{409. doi: 10.1071/WR08172.
Hanski, I. (2000), `Extinction debt and species credit in boreal forests: modelling the consequences of di erent approaches to biodiversity conservation',
Annales Zoologici Fennici pp. 271-280.
Hu, B., Li, J., Jing, L. and Judah, A. (2014), `Improving the e ciency and accuracy of individual tree
crown delineation from high-density lidar data',
International Journal of Applied Earth Observation and Geoinformation 26, 145{15. doi: 0.1016/j.jag.2013.06.003.
Jing, L., Hu, B., Li, J. and Noland, T. (2012), `Automated delineation of individual tree crowns from
lidar data by multi-scale analysis and segmentation',
Photogrammetric Engineering & Remote Sensing 78(12), 1275{1284. doi: 10.14358/PERS.78.11.1275.
Lindenmayer, D. B., Cunningham, R. B. and Donnelly, C. F. (1997), `Decay and collapse of trees with hollows in eastern australian forests: Impacts on arboreal marsupials',
Ecological Applications 7(2), 625{641. doi: 10.2307/2269526.
Lindenmayer, D. B. and Wood, J. T. (2010), `Long-term patterns in the decay, collapse, and abundance of trees with hollows in the mountain ash (eucalyptus regnans) forests of victoria, southeastern australia',
Canadian Journal of Forest Research 40(1), 48{54. doi: 10.1139/X09-185.
Lonsdale, D., Pautasso, M. and Holdenrieder, O. (2008), `Wood-decaying fungi in the forest: conservation needs and management options',
European Journal of Forest Research 127(1), 1{22. doi: 10.1007/s10342-007-0182-6.
Lovell, J. L., Jupp, D. L. B., Newnham, G. J., Coops, N. C. and Culvenor, D. S. (2005), `Simulation study for nding optimal lidar acquisition parameters for forest height retrieval',
Forest Ecology and Management 214(1), 398{412. doi: 10.1016/j.foreco.2004.07.077.5
Mcke, W., Dek, B., Schroi , A., H. M. and Pfeifer, N. (2013), `Detection of fallen trees in forested areas using small footprint airborne laser scanning data', Canadian Journal of Remote Sensing 139(s1), S32{S40. doi:10.5589/m13-013.
Miltiadou, M., Campbell, N. D., Gonzalez Aracil, S., Brown, T. and Grant, M. G. (2018), `Detection
of dead standing eucalyptus camaldulensis without tree delineation for managing biodiversity in native australian forest', International Journal of Applied Earth Observation and Geoinformation 67, 135{147. doi: https://doi.org/10.1016/j.jag.2018.01.008.
Miltiadou, M., Agapiou, A., Gonzalez Aracil, S., & Hadjimitsis, D. G. (2020). Detecting Dead Standing Eucalypt Trees from Voxelised Full-Waveform Lidar Using Multi-Scale 3D-Windows for Tackling Height and Size Variations. Forests, 11(2), 161. doi: https://doi.org/10.3390/f11020161
New South Wales Government (2016), Biodiversity conservation act 2016 no 63, Technical report.
Pasher, J. and King, D. J. (2009), `Mapping dead wood distribution in a temperate hardwood for-
est using high resolution airborne imagery',
Forest Ecology and Management 258(7), 1536{1548. doi:10.1016/j.foreco.2009.07.009.
Peterson, G., Allen, C. R. and Holling, C. S. (1998), `Ecological resilience, biodiversity, and scale',
Ecosystems 1(1), 6{18. doi: 10.1007/s100219900002.
Polewski, P., Yao, W., Heurich, M., Krzystek, P. and Stilla, U. (2015), `Detection of fallen trees in als point clouds using a normalized cut approach trained by simulation', I
SPRS Journal of Photogrammetry and Remote Sensing 105, 252{271. doi: 10.1016/j.isprsjprs.2015.01.010.
Popescu, S. C., Wynne, R. H. and Nelson, R. F. (2003), `Measuring individual tree crown diameter with lidar and assessing its in uence on estimating forest volume and biomass',
Canadian journal of remote sensing 29(5), 564{577. doi: 10.5589/m03-027.
Popescu, S. C. and Zhao, K. (2008), `A voxel-based lidar method for estimating crown base height for deciduous and pine trees', Remote sensing of environment 112(3), 767{781. doi: 10.1016/j.rse.2007.06.011.
Shendryk, I., Broich, M., Tulbure, M. G. and Alexandrov, S. V. (2016), `Bottom-up delineation of individual trees from full-waveform airborne laser scans in a structurally complex eucalypt forest',
Remote Sensing of Environment 173, 69{83. doi: 10.1016/j.rse.2015.11.008.
Shendryk, I., Broich, M., Tulbure, M. G., McGrath, A., Keith, D. and Alexandrov, S. V. (2016), `Mapping individual tree health using full-waveform airborne laser scans and imaging spectroscopy: A case study for a foodplain eucalypt forest',
Remote Sensing of Environment 187, 202{217. doi: 10.1016/j.rse.2016.10.014.
Siitonen, J. (2001), `Forest management, coarse woody debris and saproxylic organisms: Fennoscandian boreal forests as an example',
Ecological bulletins pp. 11{41.
