An early exploration of the use of the Microsoft Azure Kinect for estimation of urban tree Diameter at Breast Height

ABSTRACT

Forest and urban tree inventory measurements are increasingly adopting Remote Sensing (RS) techniques due to the accurate and rapid estimates available compared to conventional methods. The focus of this study is to assess the accuracy and potential application of the Microsoft Azure Kinect – a lightweight depth sensor – for outdoor measurement of tree stem Diameter at Breast Height (DBH). Individual urban trees (n = 51) were recorded from one viewing angle at a distance of 1 m to 5 m away using the various Field of View (FOV) settings on the depth sensor, from which resultant point clouds provided DBH estimates using a circle-fitting approach. The optimal capture method was observed at a distance of 2 m using the binned Near Field of View (NFOV) setting. Root Mean Square Error (RMSE) of DBH using this method was 8.43 cm; however, after removing trees with irregular or non-circular stems, this improved to 3.53 cm. Variations in ambient light were observed to have little effect on DBH estimates. The results of this study suggest when in an outdoor environment, the Azure Kinect should be used at a distance no greater than 3 m away, using the binned NFOV sensor setting, for DBH estimates.     

Tree hollow development in wet and dry sclerophyll eucalypt forest in south-east Queensland,Australia

Tree hollow development was studied in Eucalyptus pilularis (blackbutt) and Eucalyptus microcorys (tallowwood) in wet sclerophyll forests, and Eucalyptus signata (scribbly gum) in dry sclerophyll forest in south-east Queensland, Australia. In all three tree species hollow development appeared to be primarily initiated by fungal infection entering trees through dead branches and/or branch stubs rather than from fire scars. The process of hollow development appeared to commence when the trees were around 100 years old.

The size and abundance of hollows increased with the size and age of the tree but there were apparent differences between species. Hollows were found in the majority (>50%) of E. pilularis trees above 100 cm dbh. By contrast most E. microcorys and E. signata larger than 80 cm dbh had hollows. There was a similar difference between species in the numbers of hollows present in trees of different size. E. pilularis trees 100–110 cm dbh had two hollows per tree but this number were present in much smaller E. microcorys or E. signata trees. The rate of growth of E. pilularis is faster than either E. microcorys or E. signata and the rate of hollow formation appears to be dependent on the growth rate of the species. This suggests that hollows begin to develop after comparable times in the different species although the time taken for large hollow development differs between the species.

The analyses suggest that the smallest arboreal marsupials and hollow dependent birds could recolonise regrowth E. pilularis trees after about 165 years but hollows suitable for larger animals would not be available for around 210 years. This time frame would be longer for slower growing species of trees. Some of the implications of this work in connection with management practices that aim to ensure the maintenance of hollow-dependent fauna in these forests are discussed.     

Occurrence of potential nest trees and their use by birds in sclerophyll forest in north-east Tasmania

Breeding of hollow-nesting birds and the distribution and availability of tree hollows were investigated in an area of sclerophyll forest in north-eastern Tasmania. The occurrence of potential tree hollows was assessed from the ground and used as an index of actual hollows. Entrance size was categorized as small, medium or large corresponding to the differing requirements of the birds present on the site. Smaller hollows were more abundant than larger hollows. The distribution of hollows amongst trees was highly clumped. Diameter at breast height and the presence of a fire scar explained one-third of the variation in numbers of hollows per tree. Streamside and non-streamside areas did not differ in the number of hollows/tree or the sizes of hollows present. Thirty trees were found that were utilized as nest sites by striated pardalotes Pardalotus striatus, laughing kookaburras Dacelo novae- gaineae, green rosellas Platycercus caledonicus and yellow-tailed black cockatoo Calyptorhynchus funereus. In comparison with a representative sample of trees, nest trees were larger, had a greater number of potential hollows, a greater proportion of larger hollows and a greater proportion were fire scarred. Eucalyptus viminalis was used as a nest tree to a greater extent than expected from its relative abundance. Hollows appeared to be far in excess of the requirements of the birds present. Trees remaining in streamside reserves after logging would probably provide enough hollows to support the number of hollow-nesting birds present before logging. However, only 13% of striated pardalotes and none of the other species utilized streamside areas for nesting in the undisturbed forest.     

Variation in oil content and tree size between six geographically separate Santalum spicatum families,established near Narrogin,Western Australia

Six geographically separate sandalwood (Santalum spicatum) tree families were established together in a trial near Narrogin, Western Australia to compare oil content and tree size variation at ages 10 and 18 years. The trial site contained a total of 300 sandalwood trees with 41–55 trees from each family. All 300 sandalwood trees were measured for tree size: height, stem diameter and bole length. Wood core samples (heartwood + sapwood) were also taken at 300 mm above the ground from 71 to 72 separate trees, at ages 10 and 18 years. Each wood core sample was ground separately and then a sub-sample was analysed for oil concentration, and α- and β-santalol concentration within the oil.

The mean extractable oil concentrations within the sandalwood stems at 300 mm above the ground were significantly greater from the Borden and Dumbleyung families (1.1–1.6%) than the Ravensthorpe family (0.5–0.8%), at both ages 10 and 18 years. However, the oil quality (mean concentration of α- and β-santalol) was not significantly different between the six families at ages 10 or 18 years. Instead, oil quality was highly variable within each family.

Between sandalwood ages 10 and 18 years, the overall mean extractable oil concentration from the six different families increased significantly from 0.8 to 1.3%. During the same time period, mean α-santalol concentration increased significantly from 8.6 to 13.8%, and mean β-santalol concentration also increased significantly from 3.3 to 4.5%.

Mean tree height, stem diameter and bole length varied significantly between the six families at age 18 years. The Borden family appeared to be more tree-like in form, with a significantly longer bole length, than some of the other families which were more shrub-like in appearance.     

Estimation of forest stand diameter class using airborne lidar and field data

In this study, a method for estimating the stand diameter at breast height (DBH) classes in a South Korea forest using airborne lidar and field data was proposed. First, a digital surface model (DSM) and digital terrain model (DTM) were generated from the lidar data that have a point density of 4.3 points/m², then a tree canopy model (TCM) was created by subtracting the DTM from the DSM. The tree height and crown diameter were estimated from the rasterized TCM using local maximum points, minimum points and a circle fitting algorithm. Individual tree heights and crown diameters were converted into DBH using the allometric equations obtained from the field survey data. We calculated the proportion of the total number of individual trees belonging to each DBH class in each stand to determine the stand DBH class according to the standard guidelines. More than 60% of the stand DBH classes were correctly estimated by the proposed method, and their area occupied over 80% of the total forest area. The proposed method generated more accurate results compared to the digital forest type map provided by the government.     

Evaluating wood volume estimates derived from Quickbird imagery with GEOBIA for Pinus nigra trees in the Pentalofo forest,northern Greece

The objective of this research was to evaluate wood volume estimates of Pinus nigra trees in forest stands, which were derived utilizing Geographic Object-Based Image Analysis. Information on forest parameters such as wood volume and number of trees is useful for forest management facilitating forest sustainability. Most of the existing approaches used to estimate wood volume of forest trees require field measurements, which are laboursome. In this study, the collected field data were utilized only in order to investigate the results. Wood volume was estimated based on an individual tree crown approach and using monoscopic satellite images in combination with allometric data. The study area is the Pentalofo forest, which is located in Kozani prefecture in western Macedonia, Northern Greece. About 1 plot surface of 0.1143 ha was utilized. During the preprocessing, a pansharpened image was produced from two Quickbird satellite images (one multispectral image of 2.4 m spatial resolution and one panchromatic image of 0.6 m spatial resolution). Bands of this image were utilized single or in combination in order to delineate the tree crowns individually. The allometric equation served in order to calculate the tree Diameter at Breast Height (DBH) utilizing the detected tree crowns. The evaluation was conducted on three levels: (i) number of trees, (ii) DBH class distribution and (iii) wood volume. On the third level, the evaluation procedure was conducted twice; once using field height and once without. The difference between the results and the field data for the wood volume reached a maximum of approximately 30%. The total number of trees was exactly the same as counted in the field and the DBH distribution showed a tendency for the trees to move to a higher DBH class, resulting in an overestimation of the wood volume.     

Automated tree crown detection and size estimation using multi-scale analysis of high-resolution satellite imagery

We tested an automated multi-scale approach for detecting individual trees and estimating tree crown geometry using high spatial resolution satellite imagery. Individual tree crowns are identified as local extrema points in the Laplacian-of-Gaussian scale-space pyramid that is constructed based on linear scale-space theory. The approach simultaneously detects tree crown centres and estimates tree crown sizes (radiuses). We evaluated our method using two 0.6-m resolution QuickBird images of a forest site that underwent a large shift in tree density between image captures due to drought-associated mortality. The automated multi-scale approach produced tree count estimates with an accuracy of 54% and 73% corresponding to the dense and sparse forests, respectively. Estimated crown diameters were linearly correlated with field-measured crown diameters (r = 0.73–0.86). Tree count accuracies and size estimates were comparable with alternative methods. Future use of the presented approach is merited based on the results of our study, but requires further investigation in a broader range of forest types.     

Effects of tree spacing on branch-size development during early growth of an experimental plantation of Eucalyptus pilularis in subtropical Australia

ABSTRACT

Branch development in trees 1.6–5.7 years after planting was studied in a plantation experiment involving Eucalyptus pilularis in subtropical eastern Australia. The experiment compared stocking densities at planting in the range of 816–1667 stems ha^?1 and rectangularities of tree spacing in the range of 1–6. Branch diameters at their bases were measured on the lowest 5 m of the stems of trees judged likely to yield sawn timber of high quality at final harvest. Neither stocking density nor rectangularity had substantial effects on branch numbers or diameters. Particular interest was paid to the presence of branches with diameters in excess of 2.5 cm that might produce knots large enough to degrade the quality of sawn timber. Such branches developed on lower stems between 2–4 years of age. An average of nearly five such branches was found on more widely spaced trees, reducing to 2–3 on more closely spaced trees. The results confirmed other work suggesting that pruning of high-quality eucalypt plantations should start near 2–3 years of age to both limit the development of large branches and restrict the size of knotty cores in logs. There was no evidence that changing the rectangularity of spacing to 3 or a little more, from the more normal practice of close to square, would affect such pruning regimes: increasing the distance between rows can reduce establishment and harvest costs and allow easier access for silvicultural operations.     

Contrasting water use patterns of two important agroforestry tree species in the Mt Elgon region of Uganda

ABSTRACT

Lack of information on water use of key agroforestry species is an obstacle to understanding their influence on crop productivity. Cordia africana and Albizia coriaria are the dominant tree species of smallholder farming systems in the Mt Elgon region of Uganda and have multiple uses in agroforestry systems. This study deployed six sap flow meters on stems of three selected trees each of C. africana and A. coriaria on-farm. The objective of the study was to assess the daily water use patterns of these agroforestry tree species at different times of the year. We measured the daily sap flow of these two species using the heat ratio method over a period of 18 months. There was a significant main effect of the interaction between tree species and season on daily water use. The two species show contrasting patterns of seasonal water use across leaf shedding stages characterised by episodes of reverse flow in A. coriaria at specific periods of the year. We propose that reverse flows in A. coriaria were triggered by leaf shading while the zero flows in C. africana, which occurred during rainfall events, could have resulted from a lag phase, an indication that the two species may have different water-use strategies. Although C. africana uses 12–15 l day^?1 and A. coriaria uses 20–32 l day^?1 based on the study trees, C. africana generally uses 12% more water than A. coriaria on a standardised daily basis. Albizia coriaria exhibited radial variation of sap velocities between the inner and outer thermocouples at different periods of measurement, a phenomenon worth investigating further. The leaf shedding patterns of the two trees provide an opportunity for maximising the temporal complementarities of agroforestry systems where these trees exist. This knowledge of C. africana and A. coriaria tree water use provides critical insight for developing successful long-term tree monitoring and management programs in agroforestry systems.     

Differences between wildfire and clearfelling on the structure of montane ash forests of Victoria and their implications for fauna dependent on tree hollows

Some differences between the effects of clearfelling and wildfire on the vegetation structure of montane ash forests in the Central Highlands of Victoria are reported. The significance of these differences is considered in terms of their possible impacts on wildlife, particularly those species which are dependent on hollows in trees.

Distinct differences between the vegetation of montane ash forests which have been clearfelled and those burnt by wildfire include: –(1) the range in ages of trees in a stand, and, (2) the spatial pattern, abundance and longevity of hollow-bearing trees. In stands burnt by wildfire, the longevity of hollow-bearing trees may be more than 50 years and several forms of such trees can be present. These trees are removed from clearfelled forest. In unlogged stands and stands regenerating from selective logging the spacing of hollow-bearing trees was typically regular in distribution. In contrast, the spatial distribution of hollow-bearing trees was clustered in stands which had regenerated after clearfelling.

These differences may have a significant detrimental impact on wildlife and lead to a reduction in the abundance of some vertebrates in a given area of forest. Species such as arboreal marsupials, owls and cockatoos, which are reliant on a hollow-bearing trees for shelter and breeding, will probably be those most affected.     

The effect of TLS point cloud sampling on tree detection and diameter measurement accuracy

Terrestrial laser scanning (TLS) is a vital technique for collecting data with millimetre-level detail from the surrounding forest or individual trees. The downside of the technique is the vast amount of data it can collect, which requires efficient data processing methods. An increasing number of manual and automatic methods have been developed in recent years and most of them use point cloud sampling to fasten the processing. The challenge in the sampling procedures is the point density of TLS, which decreases rapidly as a function of distance from the scanner location. Therefore, traditional sampling procedures are not suitable for processing TLS point clouds. The present study focuses on evaluating two sampling procedures (presented in Puttonen et al. 2013) that aim to reduce the point density without losing the characteristics of the full point cloud. The study goal was to assess the effect of these two sampling procedures in tree detection and diameter at breast height (1.3 m, DBH) measurement accuracies in two sample cases. The results demonstrated that the point cloud sampling could be effectively used without losing accuracy in tree detection. However, for measurement of single-tree attributes, the use of full point cloud is recommended. Further evaluation of the methods is required with more diverse data set and due to the manual processing applied here. The automatic approach is mandatory if the approach is to be considered for more operational use.     

Polynomial taper model for Queensland plantation hoop pine

A stem taper model, involving a high order polynomial developed using plantation hoop pine (Araucaria cunninghamii Ait. ex D. Don) trees grown in Queensland, is presented. The stem profile for each tree was initially modelled using functions of diameter under bark and height, as the dependent and independent variables respectively. Parameter estimates from these individual tree regressions were subsequently modelled, using total height and diameter at breast height under bark (DBHub) as the independent variables. Finally, total height and DBHub were modelled in terms of predominant height' and diameter at breast height over bark (DBHob).

The largest mean error in the estimates of stem diameter predicted using the final model, with predominant height and DBHob as the independent variables, was 0.12 cm. The overall mean relative error in the estimate of volume, based on coefficients modelled using predominant height and DBHob, was 0.20 per cent. Apart from its accuracy, the main advantages of the model include its simplicity and its ability to model the whole tree using a single function.     

Forest Structure and Community Composition of Cochin Mangroves,South-West Coast of India

This study presents the structural characteristics of Cochin mangroves in Kerala for proposing suitable management and rejuvenation measures of degrading mangrove habitats. The floristic diversity of mangroves revealed 14 species of true mangroves belonging to six families. Multivariate analysis of true mangroves belonging to selected mangrove forests of the study area based on density could be classified into four floristic groups, a water front or low–tide zonation, mid-tide preferring species, high tide and landward zonation. Shannon Weiner index of the three stations revealed that Site I was having higher value (H′ = 2.66) followed by Site II (H′ = 2.01) and Site III (H′ = 1.595). The density of the mangroves varied significantly with sites and species (Global R = 0.537, P < 0.001). The diameter at breast height (DBH) in the study area revealed that most of the species came under 1–10 cm DBH class. The overall structural data (including Importance Value Index, DBH and basal area) showed that Site III, the Mangalavanam forest was having more structural development and could be considered as matured forest whereas, Site I, Aroor is a maturing forest and Site II, Malippuram is the least matured forest from the study. The Importance Value Index and basal area of each species could be used for analysing the maturity of the forest and habitat preferences for restoration programmes of the degraded ecosystems.     

A refined method for estimating capsule crops in individual jarrah (Eucalyptus marginata) crowns

An accurate measure of the number of capsules in the crowns of jarrah (Eucalyptus marginata) trees is needed to assess the potential for seedling regeneration prior to silvicultural treatment in jarrah forests. The current method of estimating capsule crops on jarrah trees uses stem diameter and estimates of capsule density in the crown, but has not been fully validated. In this study, we sought to develop an accurate and practical method of assessing capsule crops in the crowns of individual jarrah trees. We did this by measuring a number of tree characteristics prior to felling them. A total of 24 trees were selected, spanning a range of sizes and crown conditions, and the total number of capsules was counted for each tree. Multiple linear regression was used to model capsule number against various combinations of eight different tree/crown variables, with model fit compared using an adjusted coefficient of determination (adjR²). The final model recommended for field use included three easily measured variables (stem diameter, subjective assessment of capsule density, and subjective assessment of capsule clump distribution in the crown) and had a high degree of predictability (adjR² = 0.83), which was the same as that of the full model. This method substantially improved estimates of crown capsule numbers compared with the method currently used (adjR² increased from 0.29 to 0.83), which tended to underestimate canopy capsule numbers.     

Long-term estimates of live above-ground tree carbon stocks and net change in managed uneven-aged mixed species forests of sub-tropical Queensland,Australia

Estimates of carbon stocks and their annual change for extensive Australian sub-tropical forests are based on indirect estimates or on data derived from temperate forests. We estimated live above-ground tree carbon (LAC) stocks at landscape level from 355?000 measurements of 94?127 tree stems from 604 permanently monitored plots representing 2.6 million ha of managed uneven-aged mixed-species native forests in sub-tropical Queensland. These plots were established between 1936 and 1998 and re-measured every 2 to 10 years up to 2011. Landscapes were represented by 16 broad vegetation groups growing across a mean annual rainfall range of 500 to 2000 mm. Landscape-mean LAC stocks varied from 20.8 ± 4.3 t C ha^?1 in inland eucalypt woodlands to 146.4 ± 11.1 t C ha^?1 in coastal wet tall open forests. Landscape maximum LAC stock, defined as the mean of maximum LAC stocks over the entire measurement history for a specified landscape under prevailing environmental conditions and disturbance regimes, including sustainable forest management, ranged from 34.0 ± 7.2 t C ha^?1 in inland eucalypt woodlands to 154.9 ± 19.4 t C ha^?1 in coastal wet tall open forests. Annual live above-ground net carbon flux (C-flux) across all forests types ranged from 0.46 to 2.92 t C ha^?1 y^?1 with an overall mean of 0.95 t C ha^?1 y^?1 (n = 2067). Comparison of our results with Intergovernmental Panel on Climate Change (IPCC) estimates shows that in all cases, except for the sub-tropical steppe, the IPCC over-estimated stocks by between 13% and 34%. Conversely, the IPCC estimated C-fluxes were between 14% and 40% less than the Queensland estimates. These results extend statistically valid estimates of landscape LAC stocks and fluxes to the sub-tropical regions of Australia.     

Stem biomass estimation based on stem reconstruction from terrestrial laser scanning point clouds

Forest biomass is often difficult to quantify because field measurements are time consuming and require destructive sampling. This study explores the retrieval of stem biomass of individual trees by terrestrial laser scanning (TLS). Destructive sampling was done to collect biomass data from sample trees and used as a dependent variable in a regression analysis. Two biomass estimation models were investigated: one based on diameter at breast height (DBH) and another based on the sum of the stem section volume. Both the DBH and the stem section volume were determined from automatic reconstruction of the stem curves. Two tree species (Scots pine and Norway spruce) were considered together. The quality of the performance of the models was evaluated via a leave-one-out cross-validation strategy using accurate field measurements for 30 trees. The correlation coefficient (r) and root mean square errors (RMSEs) between the predicted and measured stem biomass were used as measures of goodness of model fitting. The model with DBH as the predictor produced an r-value of 0.93 and an RMSE of 21.5%. For the model using the reconstructed stem and correspondingly derived stem volume as the predictor, an r-value of 0.98 and an RMSE of 12.5% were achieved. The results indicated that TLS measurements are capable of assessing stem biomass with high automation and accuracy by reconstructing the stem from TLS point clouds.     

Epigaeic beetle communities of the Green Triangle plantation landscape,southern Australia: effect of remnant size,vegetation quality and structural complexity

Across the Green Triangle plantation landscape of south-eastern South Australia and south-western Victoria there remain remnant patches of native vegetation which vary considerably in both size and vegetation quality. This study focused on how the vegetation quality and or size of remnants embedded within the bluegum (Eucalyptus globulus) plantation landscape influence the local ground-dwelling Coleoptera (beetle) assemblages found within them. The aim was to particularise those remnant characteristics (in terms of size and vegetation quality) that are likely to result in the most effective protective management of local remnant beetle biodiversity. Pitfall sampling was carried out over a 13-month period. All beetles sampled via pitfall traps were identified to family level, and the three most abundant beetle families (Carabidae, Staphylinidae and Tenebrionidae) identified to genus. Remnant vegetation was assigned to one of the four quality categories (as defined by the Land for Wildlife (LFW) method), and structural attributes of the vegetation surrounding each trap were also measured.

Different beetle taxa responded independently to remnant vegetation quality and size, and to the measured vegetation structural variables. Mean beetle family abundance and estimated richness increased with decreasing remnant vegetation quality. However, when Carabidae, Staphylinidae and Tenebrionidae beetle families were analysed separately at genus level, taxa responded independently to vegetation quality and size. Furthermore, different vegetation structural variables such as the percentage cover of litter and shrubs as well as the number of overstorey stems were correlated with estimated richness of different beetle taxa. Beetle abundance was significantly influenced by mean litter depth and shrub and litter cover. This study also showed that remnants of different vegetation quality supported beetle assemblages composed of significantly different communities. However, these differences were not consistent between all sites of different vegetation quality, suggesting that there may be additional ecological characteristics within the landscape responsible for shaping remnant beetle communities. No single, specific quality of remnant vegetation was optimal for supporting the most diverse assemblage of beetles, as different taxa responded independently. A range of remnants of differing qualities maintained across the landscape may provide the resources needed for the widest range of beetle taxa, by accommodating their varying habitat requirements.     

Development of an automated individual tree detection model using point cloud LiDAR data for accurate tree counts in a Pinus radiata plantation

We develop and evaluate a new individual tree detection (ITD) algorithm to automatically locate and estimate the number of individual trees within a Pinus radiata plantation from relatively sparse airborne LiDAR point cloud data. The area of interest comprised stands covering a range of age classes and stocking levels. Our approach is based on local maxima (LM) filtering that tackles the issue of selecting the optimal search radius from the LiDAR point cloud for every potential LM using metrics derived from local neighbourhood data points; thus, it adapts to the local conditions, irrespective of canopy variability. This was achieved through two steps: (i) logistic regression model development using simulated stands composed of individual trees derived from real LiDAR point cloud data and (ii) application testing of the model using real plantation LiDAR point cloud data and geolocated, tree-level reference crowns that were manually identified in the LiDAR imagery. Our ITD algorithm performed well compared with previous studies, producing RMSE of 5.7% and a bias of only ?2.4%. Finally, we suggest that the ITD algorithm can be used for accurately estimating stocking and tree mapping, which in turn could be used to derive the plot-level metrics for an area-based approach for enhancing estimates of stand-level inventory attributes based on plot imputation.     

Estimating plot-level tree structure in a deciduous forest by combining allometric equations,spatial wavelet analysis and airborne LiDAR

Object-oriented classification methods are increasingly used to derive plant-level structural information from high-resolution remotely sensed data from plant canopies. However, many automated, object-based classification approaches perform poorly in deciduous forests compared with coniferous forests. Here, we test the performance of the automated spatial wavelet analysis (SWA) algorithm for estimating plot-level canopy structure characteristics from a light detection and ranging (LiDAR) data set obtained from a northern mixed deciduous forest. Plot-level SWA-derived and co-located ground-based measurements of tree diameter at breast height (DBH) were linearly correlated when canopy cover was low (correlation coefficient (r)?=?0.80) or moderate (r?=?0.68), but were statistically unrelated when canopy cover was high. SWA-estimated crown diameters were not significantly correlated with allometrically based estimates of crown diameter. Our results show that, when combined with allometric equations, SWA can be useful for estimating deciduous forest structure information from LiDAR in forests with low to moderate (<175% projected canopy area/ground area) levels of canopy cover.