Statistical analysis of LiDAR-derived structure and intensity variables for tree species identification (original) (raw)
Related papers
Cool temperate rainforest and adjacent forestsclassification using airborne LiDAR data
Area, 2011
The traditional methods of forest classification, based on the interpretation of aerial photographs and processing of multi-spectral and/or hyper-spectral remote sensing data are limited in their ability to capture the structural complexity of the forests compared with analysis of airborne LiDAR (light detection and ranging) data. This is because of LiDAR's penetration of forest canopies such that detailed and three-dimensional forest structure descriptions can be derived. This study applied airborne LiDAR data for the classification of cool temperate rainforest and adjacent forests in the Strzelecki Ranges, Victoria, Australia. Using normalised LiDAR point data, the forest vertical structure was stratified into three layers. Variables characterising the height distribution and density of forest components were derived from LiDAR data within each of these layers. The statistical analyses, which included one-way analysis of variance with post hoc tests, identified effective variables for forest-type classifications. The results showed that using linear discriminant analysis, an overall classification accuracy of 91.4% (as verified by the cross-validation) was achieved in the study area.
Classifying tree species using structure and spectral data from LIDAR
Two airborne laser scanning datasets with leaf-on and leaf-off conditions were used to compare parameters derived from crown structure metrics and intensity data. Five deciduous species and six coniferous species were collected at the Washington Park Arboretum, Seattle, Washington, USA. Linear (LDA) and quadratic (QDA) discriminate functions were used to classify these selected species groups. Overall, classification accuracy was highest when using intensity variables with the leaf-off data in both LDA (98.9%) and QDA (99.0%). In terms of structure variables, leaf-on variables showed higher accuracy (74.9 %) than leaf-off variables (50.2 %) while in terms of intensity variables, leaf-off variables showed higher accuracy (97.1 %) than leaf-on variables (63.0 %) in LDA. QDA showed higher classification accuracy than LDA for all cases. The overall result indicates that parameters computed from LiDAR-based crown structures and intensity data can be used to differentiate species groups and also implies that tree species classification depends on the collected LiDAR datasets and the derived parameters.
Using classification trees to predict forest structure types from LiDAR data
This study assesses whether metrics extracted from airborne Li-DAR (Light Detection and Ranging) raw point cloud can be exploited to predict different forest structure types by means of classification trees. Preliminarily, a bivariate analysis by means of Pearson statistical test was developed to find associations between LiDAR metrics and the proportion of basal area into three stem diameter classes (understory, mid-story, and over-story trees) of 243 random distributed plots surveyed from 2007 to 2012 in Trento Province (Northern Italy). An unsupervised clustering approach was adopted to determine forest structural patterns on the basis of basal area proportion in the three stem diameter classes, using a k-means procedure combined with a previous hierarchical classification algorithm. A comparison among the identified clusters centroids was performed by the Kruskall-Wallis test. A classification tree model to predict forest structural patterns originating from the cluster analysis was developed and validated. Between 18 potential LiDAR metrics, 11 were significantly correlated with the proportion of basal area of understory, mid-story, and overstory trees. The results coming from the agglomerative hierarchical clustering allowed identification of 5 clusters of forest structure: pole-stage (70% of the considered cases), young (15%), adult (24.3%), mature (24.3%), and old forests (30%). Five LiDAR metrics were selected by the classification tree to predict the forest structural types: standard deviation and mode of canopy heights, height at which 95% and 99% of canopy heights fall below, difference between height at which 90% and 10% of canopy heights fall below. The validation tree model process showed a misclassification error of 45.9% and a level of user's accuracy ranging between 100% and 33.3% in the validation data set. The highest level of user's accuracy was reached in the classification of pole-stage forests (100%), in which more than 82% of basal area is due to the understory-trees, follow by the classification of old forests types (63.5% of basal area due to the overstory-trees) achieved 76.5% of user's accuracy. The model has provided moderately satisfactory results in term of classification performance: substantial room for improvement might be established by multi-or hyperspectral imaging that allow detailed characterization of the spectral behaviour of the forest structure types. Keywords airborne laser scanning, discrete return laser scanner data, stem diameter classes, basal area, bivariate analysis, unsupervised clustering, classification tree model, forest inventory, forest management
Isprs Journal of Photogrammetry and Remote Sensing, 2018
Light detection and ranging (lidar) data have been increasingly used for forest classification due to its ability to penetrate the forest canopy and provide detail about the structure of the lower strata. In this study we demonstrate forest classification approaches using airborne lidar data as inputs to random forest and linear unmixing classification algorithms. Our results demonstrated that both random forest and linear unmixing models identified a distribution of rainforest and eucalypt stands that was comparable to existing ecological vegetation class (EVC) maps based primarily on manual interpretation of high resolution aerial imagery. Rainforest stands were also identified in the region that have not previously been identified in the EVC maps. The transition between stand types was better characterised by the random forest modelling approach. In contrast, the linear unmixing model placed greater emphasis on field plots selected as endmembers which may not have captured the variability in stand structure within a single stand type. The random forest model had the highest overall accuracy (84%) and Cohen's kappa coefficient (0.62). However, the classification accuracy was only marginally better than linear unmixing. The random forest model was applied to a region in the Central Highlands of southeastern Australia to produce maps of stand type probability, including areas of transition (the 'ecotone') between rainforest and eucalypt forest. The resulting map provided a detailed delineation of forest classes, which specifically recognised the coalescing of stand types at the landscape scale. This represents a key step towards mapping the structural and spatial complexity of these ecosystems, which is important for both their management and conservation.
LIDAR remote sensing for secondary Tropical Dry Forest identification
2012
This study evaluates the use of waveform LIDAR data for the characterization of secondary forests stages in a Tropical Dry Forest (TDF) area of Guanacaste, Costa Rica. A secondary forest succession is defined here as regrowth of woody vegetation following a complete or heavy forest clearance for pasture, agriculture, or other human activity. We first compare the known spatial distribution of three main TDF successional stages (Early, Intermediate and Late successional stages) as constrained by published field observations of tree height with that obtained from a three-class classification of the LIDAR data. In doing so we explicitly assess the possibility of using LIDAR data to map the distribution of the three main TDFs stages. We demonstrates that changes in the forest vertical structure (such as height) associated with principal successional stages (Early, Intermediate and Late) of TDF secondary growth can be effectively identified from LIDAR data. The successional sequence observed is related to changes in the vertical distribution of woody components that occur when forest patches evolve from an Early Stage which is dominated by sparse trees and grass, to initial stages of intermediate secondary succession characterized by small canopies and a high density of understory regeneration, and then to more complex and developed stages of intermediate and late stage secondary succession which share some attributes of Old growth forests. We then focus on the Intermediate successional Stage since it is of highest interest to policy makers dealing with programs implementing payments for environmental services. We show that an analysis of the LIDAR data can identify successional three subclasses within the Intermediate Stage providing further insights in the development of secondary forest growth. A validation of the three new forest classes is provided using field observations showing a mean tree height and standard deviation of 6.16 ± 0.87, 7.82 ± 0.31 and 8.62 ± 1.22. m.
Assessing forest metrics with a ground-based scanning lidar
Canadian Journal of Forest Research-revue Canadienne De Recherche Forestiere, 2004
A ground-based scanning lidar (light detection and ranging) system was evaluated to assess its potential utility for tree-level forest mensuration data extraction. Ground-based-lidar and field-mensuration data were collected for two forest plots: one located within a red pine (Pinus resinosa Ait.) plantation and another in a mixed deciduous stand dominated by sugar maple (Acer saccharum Marsh.). Five lidar point cloud scans were collected from different vantage points for each plot over a 6-h period on 5 July 2002 using an Optech Inc. ILRIS-3D laser imager. Fieldvalidation data were collected manually over several days during the same time period. Parameters that were measured in the field or derived from manual field measures included (i) stem location, (ii) tree height, (iii) stem diameter at breast height (DBH), (iv) stem density, and (v) timber volume. These measures were then compared with those derived from the ILRIS-3D data (i.e., the lidar point cloud data). It was found that all parameters could be measured or derived from the data collected by the ground-based lidar system. There was a slight systematic underestimation of mean tree height resulting from canopy shadow effects and suboptimal scan sampling distribution. Timber volume estimates for both plots were within 7% of manually derived estimates. Tree height and DBH parameters have the potential for objective measurement or derivation with little manual intervention. However, locating and counting trees within the lidar point cloud, particularly in the multitiered deciduous plot, required the assistance of field-validation data and some subjective interpretation. Overall, ground-based lidar demonstrates promise for objective and consistent forest metric assessment, but work is needed to refine and develop automatic feature identification and data extraction techniques.
Estimating Tropical Forest Structure Using a Terrestrial Lidar
PloS one, 2016
Forest structure comprises numerous quantifiable biometric components and characteristics, which include tree geometry and stand architecture. These structural components are important in the understanding of the past and future trajectories of these biomes. Tropical forests are often considered the most structurally complex and yet least understood of forested ecosystems. New technologies have provided novel avenues for quantifying biometric properties of forested ecosystems, one of which is LIght Detection And Ranging (lidar). This sensor can be deployed on satellite, aircraft, unmanned aerial vehicles, and terrestrial platforms. In this study we examined the efficacy of a terrestrial lidar scanner (TLS) system in a tropical forest to estimate forest structure. Our study was conducted in January 2012 at La Selva, Costa Rica at twenty locations in a predominantly undisturbed forest. At these locations we collected field measured biometric attributes using a variable plot design. We...
2011
Although remotely sensed data have been widely explored for forest applications, passive remote sensing techniques are limited in their ability to capture forest structural complexity, particularly in uneven-aged, mixed species forests with multiple canopy layers. Generally, these techniques are only able to provide information on horizontal (two-dimensional) forest extent. The vertical forest structure (or the interior of the canopy and understorey vegetation) cannot be mapped using these passive remote sensing techniques. Fortunately, it has been shown that active remote sensing techniques via airborne LiDAR (light detection and ranging) with capability of canopy penetration yields such high density sampling that detailed description of the forest structure in three-dimensions can be obtained. Accordingly, much interest is attached to exploring the application of this approach for identifying the distribution of designated vegetation communities. However, the suitability of LiDAR ...
LiDAR remote sensing of forest structure
2016
Abstract: Light detection and ranging (LiDAR) technology provides horizontal and vertical information at high spatial resolutions and vertical accuracies. Forest attributes such as canopy height can be directly retrieved from LiDAR data. Direct retrieval of canopy height provides opportunities to model above-ground biomass and canopy volume. Access to the vertical nature of forest ecosystems also offers new opportunities for enhanced forest monitoring, management and planning. Key words: biomass, forest structure, laser altimetry, LiDAR, remote sensing. I