Full-waveform LiDAR Research Articles

A Novel De-Noising Method for Improving the Performance of Full-Waveform LiDAR Using Differential Optical Path

A novel de-noising method for improving the performance of full-waveform light detection and ranging (LiDAR) based on differential optical path is proposed, and the mathematical models of this method are developed and verified. Backscattered full-waveform signal (BFWS) is detected by two avalanche photodiodes placed before and after the focus of the focusing lens. On the basis of the proposed method, some simulations are carried out and conclusions are achieved. (1) Background noise can be suppressed effectively and peak points of the BFWS are transformed into negative-going zero-crossing points as stop timing moments. (2) The relative increment percentage of the signal-to-noise ratio based on the proposed method first dramatically increases with the increase of the distance, and then the improvement gets smaller by increasing the distance. (3) The differential Gaussian fitting with the Levenberg-Marquardt algorithm is applied, and the results show that it can decompose the BFWS with high accuracy. (4) The differential distance should not be larger than c/2 × τrmin, and two variable gain amplifiers can eliminate the inconsistency of two differential beams. The results are beneficial for designing a better performance full-waveform LiDAR.

SURFACE FITTING FILTERING OF LIDAR POINT CLOUD WITH WAVEFORM INFORMATION

Abstract. Full-waveform LiDAR is an active technology of photogrammetry and remote sensing. It provides more detailed information about objects along the path of a laser pulse than discrete-return topographic LiDAR. The point cloud and waveform information with high quality can be obtained by waveform decomposition, which could make contributions to accurate filtering. The surface fitting filtering method with waveform information is proposed to present such advantage. Firstly, discrete point cloud and waveform parameters are resolved by global convergent Levenberg Marquardt decomposition. Secondly, the ground seed points are selected, of which the abnormal ones are detected by waveform parameters and robust estimation. Thirdly, the terrain surface is fitted and the height difference threshold is determined in consideration of window size and mean square error. Finally, the points are classified gradually with the rising of window size. The filtering process is finished until window size is larger than threshold. The waveform data in urban, farmland and mountain areas from “WATER (Watershed Allied Telemetry Experimental Research)” are selected for experiments. Results prove that compared with traditional method, the accuracy of point cloud filtering is further improved and the proposed method has highly practical value.

An Integrated Approach to Generating Accurate DTM from Airborne Full-Waveform LiDAR Data

In this study, full-waveform LiDAR data were exploited to detect weak returns backscattered by the bare terrain underneath vegetation canopies and thus improve the generation of a digital terrain model (DTM). Building on the methods of progressive generation of triangulation irregular network (TIN) model reported in the literature, we proposed an integrated approach where echo detection, terrain identification, and TIN generation were carried out iteratively. The proposed method was tested on a dataset collected by a Riegl LMS Q-560 scanner over a study area near Sault Ste. Marie, Ontario, Canada (46°33′56′′N, 83°25′18′′W). The results demonstrated that more terrain points under shrubs could be identified, and the generated DTMs exhibited more details in the terrain than those obtained using the progressive TIN method. In addition, 1275 points across this study area were surveyed on the ground and used to validate the proposed approach. The estimated elevations were shown to have a strong linear relationship with the measured ones, with R2 values above 0.98, and the RMSEs (Root Mean Squared Errors) between them were less than 0.15 m even for areas with hilly terrains underneath vegetation canopies.

Bayesian decomposition of full waveform LiDAR data with uncertainty analysis

A thorough understanding of full waveform (FW) LiDAR data processing and associated uncertainty is critical to vegetation applications such as retrieving forest structure variables and estimating forest biomass. This paper applies the Bayesian non-linear modeling concept to process small-footprint FW LiDAR data (the Bayesian decomposition) collected at a study site of the National Ecological Observatory Network (NEON) to investigate its potential for waveform decomposition and uncertainty estimation. Specifically, several possible models suitable for fitting waveforms were assessed within the Bayesian framework, and the Gaussian model was selected to perform the Bayesian decomposition. Subsequently, we conducted performance evaluation and uncertainty analysis at the parameter, derived point cloud and surface model levels. Results of the model reasonableness show that the Gaussian model is superior to alternative models with respect to uncertainty, physical meaning and processing efficiency. After converting waveforms to discrete points, the model comparisons demonstrate that the Bayesian decomposition can be utilized for FW LiDAR data processing, and its results are comparable to the direct decomposition (DD), Gold and RL (Richardson–Lucy) approaches in terms of the root mean squared error (RMSE<0.93m) of the point distances between the waveform-based point cloud and the reference point cloud. Additionally, more points can be extracted from FW LiDAR data with these methods than discrete-return LiDAR data, especially at the mid-story of vegetation based on the results of height bins, percentile heights and canopy LiDAR density at the individual tree level. Moreover, uncertainty estimates from the Bayesian method enhance the credibility of decomposition results in a probabilistic sense to capture the true error of estimates and trace the uncertainty propagation along the processing steps. For example, results of the surface model yield larger RMSE values (1.38m vs. 0.65m) with a wider credible interval than quantile point clouds with a more compact distribution. In contrast to commonly used deterministic approaches, the Bayesian decomposition method can produce an ensemble of reasonable parameter estimates with probability through Markov Chain Monte Carlo (MCMC) sampling from the posterior distribution of model parameters. These parameter estimates and corresponding derived products can be queried to provide meaningful interpretation of results and associated uncertainty. Both the flat priors and empirical priors can achieve good performance of the decomposition while the empirical priors tend to significantly speed up the model convergence. The Bayesian approach also renders an important insight into the uncertainty of the model performance evaluation using field data by generating reasonable prediction intervals to reduce inherent errors of field measurements.

Estimation of FPAR and FPAR profile for maize canopies using airborne LiDAR

FPAR (fraction of photosynthetically active radiation) and FPAR profile (vertical FPAR distribution) are important parameters for characterizing the vegetation growth status and studying global climate change. Few studies have been carried out to estimate FPAR and FPAR profile using waveform LiDAR data. This research explored the potential of airborne small-footprint full-waveform LiDAR in the estimation of FPAR and FPAR profile of the maize canopy in Huailai County of Hebei Province, China. First, the maize growing area was identified by a simple decision tree model. Second, raw waveform data were processed to extract LiDAR-derived energy ratio and energy ratio profile. Third, FPAR and FPAR profile were estimated from LiDAR-derived metrics. Finally, we analyzed the FPAR and FPAR profile estimation results and assessed the model validity using the leave-one-out cross-validation (LOOCV) method. The comparative analyses found that the LiDAR-derived energy ratio profile and field-measured FPAR profile had the same trend and similar change rate for all maize layers. The accuracy assessments indicated that the FPAR and FPAR profile were estimated well by the LiDAR waveform data, with the high R2 (0.90 for the whole canopy, and 0.95, 0.90, 0.93, 0.92, and 0.97 for layers 1–5) and low RMSEs (0.042 for the whole canopy, and 0.033, 0.035, 0.039, 0.043, and 0.044 for layers 1–5). The spatial distribution map of FPAR was produced to describe the maize growth status of the whole study area, and the map showed that the FPAR distributed relatively uniformly. This study suggested that airborne small-footprint full-waveform LiDAR was useful in accurately measuring FPAR and FPAR profile of the maize canopy and in effectively mapping the maize FPAR spatial distribution.

Estimation of coniferous forest aboveground biomass with aggregated airborne small-footprint LiDAR full-waveforms.

Forest aboveground biomass (AGB) is critical for assessing forest productivity and evaluating carbon sequestration rates. Discrete-return LiDAR has been widely used to estimate forest AGB, however, fewer studies have estimated the coniferous forest AGB using airborne small-footprint full-waveform LiDAR data. The objective of this study was to extract a suite of newly proposed metrics from airborne small-footprint full-waveform LiDAR data and to evaluate the ability of these metrics in estimating coniferous forest AGB. To achieve this goal, each waveform was first preprocessed, including de-noising, smoothing, and normalization. Next, all the waveforms within each plot were aggregated into a large pseudo waveform and the return energy profile was generated. Then, the foliage profile was retrieved from the return energy profile based on the Geometric Optical and Radiative Transfer (GORT) model. Finally, a series of new return energy profile metrics and foliage profile metrics were extracted to estimate forest AGB. Simple linear regression was conducted to assess the correlation between each LiDAR metric and forest AGB. Stepwise multiple regression analysis was then carried out to select important prediction metrics and establish the optimal forest AGB estimation model. Results indicated that both return energy profile and foliage profile based height-related metrics were strongly correlated to forest AGB. The energy weighted canopy height (HEweight) (R = 0.88) and foliage area weighted height (HFweight) (R = 0.89) all had the highest correlation coefficients with forest AGB in return energy profile metrics and foliage profile metrics respectively. Energy height percentiles and foliage height percentiles also had the ability to explain AGB variation. The energy-related metrics, foliage area-related metrics, and bounding volume-related metrics derived from the return energy profile and foliage profile were not all sensitive to forest AGB. This study also concluded that combining return energy profile metrics and foliage profile metrics could improve the accuracy of forest AGB estimation, and the optimal model contained the metrics of HFweight, HEweight, and VolumemaxHE, which is the product of the maximum canopy return energy profile amplitude (maxCE) and the maximum height of return energy profile (maxHE).

Simulating the Effects of the Airborne Lidar Scanning Angle, Flying Altitude, and Pulse Density for Forest Foliage Profile Retrieval

Foliage profile is a key biophysical parameter for forests. Airborne Light Detection and Ranging is an effective tool for vegetation parameter retrieval. Data acquisition conditions influence the estimation of biophysical parameters. To acquire accurate foliage profiles at the lowest cost, we used simulations to explore the effects of data acquisition conditions on forest foliage profile retrieval. First, a 3-D forest scene and the airborne small-footprint full-waveform LiDAR data were simulated by the DART model. Second, the foliage profile was estimated from LiDAR data based on a Geometric Optical and Radiative Transfer model. Lastly, the effects of the airborne LiDAR scanning angle, flying altitude, and pulse density on foliage profile retrieval were explored. The results indicated that the scanning angle was an important factor in the foliage profile retrieval, and the optimal scanning angle was 20°. The optimal scanning angle was independent of flying altitude and pulse density, and combinations of multiple scanning angles could improve the accuracy of the foliage profile estimation. The flying altitude and pulse density had little influence on foliage profile retrieval at plot level and could be ignored. In general, our study provides reliable information for selecting the optimal instrument operational parameters to acquire more accurate foliage profiles and minimize data acquisition costs.

A linearly approximated iterative Gaussian decomposition method for waveform LiDAR processing

Full-waveform LiDAR (FWL) decomposition results often act as the basis for key LiDAR-derived products, for example canopy height, biomass and carbon pool estimation, leaf area index calculation and under canopy detection. To date, the prevailing method for FWL product creation is the Gaussian Decomposition (GD) based on a non-linear Levenberg-Marquardt (LM) optimization for Gaussian node parameter estimation. GD follows a “greedy” approach that may leave weak nodes undetected, merge multiple nodes into one or separate a noisy single node into multiple ones. In this manuscript, we propose an alternative decomposition method called Linearly Approximated Iterative Gaussian Decomposition (LAIGD method). The novelty of the LAIGD method is that it follows a multi-step “slow-and-steady” iterative structure, where new Gaussian nodes are quickly discovered and adjusted using a linear fitting technique before they are forwarded for a non-linear optimization. Two experiments were conducted, one using real full-waveform data from NASA’s land, vegetation, and ice sensor (LVIS) and another using synthetic data containing different number of nodes and degrees of overlap to assess performance in variable signal complexity. LVIS data revealed considerable improvements in RMSE (44.8% lower), RSE (56.3% lower) and rRMSE (74.3% lower) values compared to the benchmark GD method. These results were further confirmed with the synthetic data. Furthermore, the proposed multi-step method reduces execution times in half, an important consideration as there are plans for global coverage with the upcoming Global Ecosystem Dynamics Investigation LiDAR sensor on the International Space Station.

Retrieval of higher order statistical moments from full-waveform LiDAR data for tree species classification

Tree inventories, inter alia, need information on the tree species. Previous studies aimed at distinguishing tree species based on three dimensional tree structure metrics derived from airborne laser scanning (ALS) point clouds or based on features from full-waveform (FW) laser scanning data provided by today's sensors. Classifications based on FW features mainly use echo amplitude, pulse energy (hereafter referred to as energy) and width, which are typically retrieved by waveform decomposition, often performed using the symmetric Gaussian distribution function.However, for forested areas, the symmetry of the echo shape is potentially modified by multiple scattering and the distribution of scattering elements (e.g. leaves). We assess the potential of processing full-waveform ALS data such that the third and fourth statistical moments, i.e. the echo skewness and the echo kurtosis, can be retrieved in addition to the amplitude, energy and FWHM. We propose a waveform decomposition approach using the skew normal distribution (SND) function, which enables the modelling of skewed echoes. We investigate the difference of tree-crown aggregated SND derived FW features between seven tree species (969 individual trees) in a temperate mixed forest with the aim of detecting the most descriptive echo features. The such derived FW features are tested for species classification.The results reveal that the largest differences across the tree species are in the mean energy of the first echoes (15 out of 21 species pairs show differences), followed by the mean amplitude of the first echoes and the mean skewness of all echoes originating from a single crown (14 out of 21 species pairs show differences against each other for both features).The differentiation of coniferous and deciduous trees benefits from the features derived from the SND decomposition compared to the use of echo amplitude only (0.39 vs. 0.61 in Cohen's κ). As the classification accuracy of the three dominant tree species within the test site only shows a small increase (0.20 vs. 0.26 in Cohen's κ) by adding FW features, we propose the use of such features in combination with features from multispectral data for this purpose.The SND decomposition is comparable to the Gaussian decomposition regarding the decomposition accuracy (RMSE=4.45 vs. RMSE=3.50) and computational cost. Hence, we propose the default use of the SND decomposition, as the SND is a more flexible function, allowing for the modelling of normally distributed echoes, as well as the fitting of skewed echoes, while no limitations regarding the direction of the skewness are introduced. We attribute the difficulties in the tree species classification to the relatively wide ranges of the crown aggregated features within one species, which for some features results in a considerable overlap of the feature ranges across the species.

Linking stand architecture with canopy reflectance to estimate vertical patterns of light-use efficiency

Remote sensing of plant carbon uptake, or gross primary production (GPP), in a repeatable and consistent manner remains a key element of a comprehensive understanding of the role of vegetation within the global carbon cycle. To further this understanding at a landscape level or global scale, accurate remote sensing of photosynthetic light-use efficiency (LUE) is required to understand photosynthetic down-regulation and environmental constraints to plant photosynthesis. The past decade has seen advances in detecting both leaf- and canopy-level physiological stress behaviours using the photochemical reflectance index (PRI), a narrow-waveband normalized difference index that relates LUE to a xanthophyll-induced absorption feature at 531nm. To date, however, much of this research has occurred using top of canopy measurements, while our understanding of the vertical distribution of LUE within the crown is limited. In this study, we demonstrate an approach which could be used to scale photosynthetic behaviour of vegetation vertically and horizontally using estimates of vertical canopy structure obtained from terrestrial Light Detection and Ranging (LiDAR) data to predict proportions of shaded and sunlit canopy which are then linked to predictions of LUE. We apply the approach over a mature Aspen study site located in central Saskatchewan, Canada utilising full-waveform LiDAR data provided by the ground-based laser scanner system and canopy spectra obtained by the AMSPEC II spectro-radiometer. Agreement between predictions of Gross Primary Productivity (GPP) using the developed approach compared to independent observations was highly significant at hourly intervals (R2=0.80, p<0.01) under clear sky conditions. A range of LUE vertical profiles for different stand structures across the growing season were developed providing estimations of how crown structure can impact LUE vertically in the crown. We conclude with a recommendation for ongoing research to verify these types of trends using concurrently acquired, independently derived leaf LUE from photosynthesis light-response curves, and forest structure variation from LiDAR, to provide a more exact quantification of these patterns.

Decomposition of LiDAR waveforms by B-spline-based modeling

Waveform decomposition is a widely used technique for extracting echoes from full-waveform LiDAR data. Most previous studies recommended the Gaussian decomposition approach, which employs the Gaussian function in laser pulse modeling. As the Gaussian-shape assumption is not always satisfied for real LiDAR waveforms, some other probability distributions (e.g., the lognormal distribution, the generalized normal distribution, and the Burr distribution) have also been introduced by researchers to fit sharply-peaked and/or heavy-tailed pulses. However, these models cannot be universally used, because they are only suitable for processing the LiDAR waveforms in particular shapes. In this paper, we present a new waveform decomposition algorithm based on the B-spline modeling technique. LiDAR waveforms are not assumed to have a priori shapes but rather are modeled by B-splines, and the shape of a received waveform is treated as the mixture of finite transmitted pulses after translation and scaling transformation. The performance of the new model was tested using two full-waveform data sets acquired by a Riegl LMS-Q680i laser scanner and an Optech Aquarius laser bathymeter, comparing with three classical waveform decomposition approaches: the Gaussian, generalized normal, and lognormal distribution-based models. The experimental results show that the B-spline model performed the best in terms of waveform fitting accuracy, while the generalized normal model yielded the worst performance in the two test data sets. Riegl waveforms have nearly Gaussian pulse shapes and were well fitted by the Gaussian mixture model, while the B-spline-based modeling algorithm produced a slightly better result by further reducing 6.4% of fitting residuals, largely benefiting from alleviating the adverse impact of the ringing effect. The pulse shapes of Optech waveforms, on the other hand, are noticeably right-skewed. The Gaussian modeling results deviated significantly from original signals, and the extracted echo parameters were clearly inaccurate and unreliable. The B-spline-based method performed significantly better than the Gaussian and lognormal models by reducing 45.5% and 11.5% of their fitting errors, respectively. Much more precise echo properties can accordingly be retrieved with a high probability. Benefiting from the flexibility of B-splines on fitting arbitrary curves, the new method has the potentiality for accurately modeling various full-waveform LiDAR data, whether they are nearly Gaussian or non-Gaussian in shape.

Total canopy transmittance estimated from small-footprint, full-waveform airborne LiDAR

Canopy transmittance is a directional and wavelength-specific physical parameter that quantifies the amount of radiation attenuated when passing through a vegetation layer. The parameter has been estimated from LiDAR data in many different ways over the years. While early LiDAR methods treated each returned echo equally or weighted the echoes according to their return order, recent methods have focused more on the echo energy. In this study, we suggest a new method of estimating the total canopy transmittance considering only the energy of ground echoes. Therefore, this method does not require assumptions for the reflectance or absorption behavior of vegetation. As the oblique looking geometry of LiDAR is explicitly considered, canopy transmittance can be derived for individual laser beams and can be mapped spatially. The method was applied on a contemporary full-waveform LiDAR data set collected under leaf-off conditions and over a study site that contains two sub regions: one with a mixed (coniferous and deciduous) forest and another that is predominantly a deciduous forest in an alluvial plain. The resulting canopy transmittance map was analyzed for both sub regions and compared to aerial photos and the well-known fractional cover method. A visual comparison with aerial photos showed that even single trees and small canopy openings are visible in the canopy transmittance map. In comparison with the fractional cover method, the canopy transmittance map showed no saturation, i.e., there was better separability between patches with different vegetation structure.

Above-ground biomass estimation using airborne discrete-return and full-waveform LiDAR data in a coniferous forest

The estimation of forest aboveground biomass (AGB) is critical for quantifying carbon stocks and essential for evaluating global carbon cycle. Many previous studies have estimated forest AGB using airborne discrete-return Light Detection and Ranging (LiDAR) data, while fewer studies predicted forest AGB using airborne full-waveform LiDAR data. The objective of this work was to evaluate the utility of airborne discrete-return and full-waveform LiDAR data in estimating forest AGB. To fulfill the objective, airborne discrete-return LiDAR-derived metrics (DR-metrics), full-waveform LiDAR-derived metrics (FW-metrics) and structure parameters (combining height metrics and canopy cover) were used to estimate forest AGB. Additionally, the combined use of DR- and FW-metrics through a nonlinear way was also evaluated for AGB estimation in a coniferous forest in Dayekou, Gansu province of China. Results indicated that both height metrics derived from discrete-return and full-waveform LiDAR data were stronger predictors of forest AGB compared with other LiDAR-derived metrics. Canopy cover derived from discrete-return LiDAR data was not sensitive to forest AGB, while canopy cover estimated by full-waveform LiDAR data (CCWF) showed moderate correlation with forest AGB. Structure parameters derived from full-waveform LiDAR data, such as H75FW*CCFW, were closely related to forest AGB. In contrast, structure parameters derived from discrete-return LiDAR data were not suitable for estimating forest AGB due to the less sensitivity of canopy cover CCDR2 to forest AGB. This research also concluded that the synergistic use of DR- and FW-metrics can provide better AGB estimates in coniferous forest.

Full-waveform LiDAR echo decomposition based on wavelet decomposition and particle swarm optimization

To measure the distances and properties of the objects within a laser footprint, a decomposition method for full-waveform light detection and ranging (LiDAR) echoes is proposed. In this method, firstly, wavelet decomposition is used to filter the noise and estimate the noise level in a full-waveform echo. Secondly, peak and inflection points of the filtered full-waveform echo are used to detect the echo components in the filtered full-waveform echo. Lastly, particle swarm optimization (PSO) is used to remove the noise-caused echo components and optimize the parameters of the most probable echo components. Simulation results show that the wavelet-decomposition-based filter is of the best improvement of SNR and decomposition success rates than Wiener and Gaussian smoothing filters. In addition, the noise level estimated using wavelet-decomposition-based filter is more accurate than those estimated using other two commonly used methods. Experiments were carried out to evaluate the proposed method that was compared with our previous method (called GS-LM for short). In experiments, a lab-build full-waveform LiDAR system was utilized to provide eight types of full-waveform echoes scattered from three objects at different distances. Experimental results show that the proposed method has higher success rates for decomposition of full-waveform echoes and more accurate parameters estimation for echo components than those of GS-LM. The proposed method based on wavelet decomposition and PSO is valid to decompose the more complicated full-waveform echoes for estimating the multi-level distances of the objects and measuring the properties of the objects in a laser footprint.

Decomposition of small-footprint full waveform LiDAR data based on generalized Gaussian model and grouping LM optimization

Full waveform airborne Light Detection And Ranging(LiDAR) data contains abundant information which may overcome some deficiencies of discrete LiDAR point cloud data provided by conventional LiDAR systems. Processing full waveform data to extract more information than coordinate values alone is of great significance for potential applications. The Levenberg–Marquardt (LM) algorithm is a traditional method used to estimate parameters of a Gaussian model when Gaussian decomposition of full waveform LiDAR data is performed. This paper employs the generalized Gaussian mixture function to fit a waveform, and proposes using the grouping LM algorithm to optimize the parameters of the function. It is shown that the grouping LM algorithm overcomes the common drawbacks which arise from the conventional LM for parameter optimization, such as the final results being influenced by the initial parameters, possible algorithm interruption caused by non-numerical elements that occurred in the Jacobian matrix, etc. The precision of the point cloud generated by the grouping LM is evaluated by comparing it with those provided by the LiDAR system and those generated by the conventional LM. Results from both simulation and real data show that the proposed algorithm can generate a higher-quality point cloud, in terms of point density and precision, and can extract other information, such as echo location, pulse width, etc., more precisely as well.

Assessment of Elevation Uncertainty in Salt Marsh Environments using Discrete-Return and Full-Waveform Lidar

ABSTRACT Rogers, J.N.; Parrish, C.E.; Ward, L.G., and Burdick, D.M., 2016. Assessment of elevation uncertainty in salt marsh environments using discrete-return and full-waveform lidar. In: Brock, J.C.; Gesch, D.B.; Parrish, C.E.; Rogers, J.N., and Wright, C.W. (eds.), Advances in Topobathymetric Mapping, Models, and Applications. Journal of Coastal Research, Special Issue, No. 76, pp. 107–122. Coconut Creek (Florida), ISSN 0749-0208. Lidar data can serve as an important source of elevation information for studying, monitoring and managing salt marshes. However, previous studies have shown that lidar data tend to have greater vertical uncertainty in salt marshes than in other environments, hindering the ability to resolve small elevation differences that can be ecologically significant in marshes. For coastal scientists and managers to effectively collect, evaluate, and/or use lidar data in salt marshes, factors affecting elevation uncertainty (e.g., plant species, season, and lidar processing methods) mus...

Improved Urban Scene Classification Using Full-Waveform Lidar

Full-waveform lidar data provides supplementary radiometric as well as more accurate geometric target information, when compared to discrete return systems. In this research, a wide range of classes in an urban scene; including trees, medium vegetation, low vegetation (grass), water bodies, pitched roofs, flat roofs, asphalt, vehicles, power lines, walls (fences) and concrete are considered. In order to tackle the challenge of distinguishing geometrically similar classes and enhancing the separability of other targets, a new set of features based on deconvolved waveforms is introduced. The positive effect of the proposed feature dataset on classification accuracy in individual classes is shown using two ensemble classifiers (random forests and RUSBoost). Performance of the classifiers is improved by integration with sampling techniques, especially for the under-represented classes. The final output of the proposed method is a highly detailed land cover map of the urban scene, which affords good separability between critical classes.

Using dual-wavelength, full-waveform airborne lidar for surface classification and vegetation characterization

This study examines the utility of cocollected, dual-wavelength, full-waveform lidar data to characterize vegetation and landscapes through the extraction of waveform features, such as total waveform energy, canopy energy distribution, and foliage penetration metrics. Assessments are performed using data collected in May 2014 over Monterey, California, using the Chiroptera dual-laser lidar mapping system from Airborne Hydrography AB. Both full-waveform and discrete return data were collected simultaneously at green (532 nm) and near-infrared (NIR) (1064 nm) wavelengths; however, the two channels are operated independently at different pulse repetition frequencies, thus measurements are not spatially coincident. A voxelization approach is employed to generate pseudowaveforms for each wavelength along vertical columns in a regularly spaced grid, such that spectral waveform properties can be evaluated independently of spatial variations resulting from instrumentation configuration and collection scenario. The pseudowaveforms are parameterized and extracted parameters are mapped to raster layers, which are then used as inputs to a random forest classifier to predict land cover classifications across the survey area. In comparison to independent classification results for the two wavelength channels, the combination of the NIR and green response provided an improvement in overall classification accuracy of up to 6%. This effort presents the methodology associated with the voxelization approach and the exploitation of the pseudowaveform features, while illustrating a potential utility for geospatial classification using multiple wavelengths.

CHP Toolkit: Case Study of LAIe Sensitivity to Discontinuity of Canopy Cover in Fruit Plantations

This paper presents an open-source canopy height profile (CHP) toolkit designed for processing small-footprint full-waveform LiDAR data to obtain the estimates of effective leaf area index (LAIe) and CHPs. The use of the toolkit is presented with a case study of LAIe estimation in discontinuous-canopy fruit plantations. The experiments are carried out in two study areas, namely, orange and almond plantations, with different percentages of canopy cover (48% and 40%, respectively). For comparison, two commonly used discrete-point LAIe estimation methods are also tested. The LiDAR LAIe values are first computed for each of the sites and each method as a whole, providing “apparent” site-level LAIe, which disregards the discontinuity of the plantations' canopies. Since the toolkit allows for the calculation of the study area LAIe at different spatial scales, between-tree-level clumping can be easily accounted for and is then used to illustrate the impact of the discontinuity of canopy cover on LAIe retrieval. The LiDAR LAIe estimates are therefore computed at smaller scales as a mean of LAIe in various grid-cell sizes, providing estimates of “actual” site-level LAIe. Subsequently, the LiDAR LAIe results are compared with theoretical models of “apparent” LAIe versus “actual” LAIe, based on known percent canopy cover in each site. The comparison of those models to LiDAR LAIe derived from the smallest grid-cell sizes against the estimates of LAIe for the whole site has shown that the LAIe estimates obtained from the CHP toolkit provided values that are closest to those of theoretical models.

A Sparsity-Based Regularization Approach for Deconvolution of Full-Waveform Airborne Lidar Data

Full-waveform lidar systems capture the complete backscattered signal from the interaction of the laser beam with targets located within the laser footprint. The resulting data have advantages over discrete return lidar, including higher accuracy of the range measurements and the possibility of retrieving additional returns from weak and overlapping pulses. In addition, radiometric characteristics of targets, e.g., target cross-section, can also be retrieved from the waveforms. However, waveform restoration and removal of the effect of the emitted system pulse from the returned waveform are critical for precise range measurement, 3D reconstruction and target cross-section extraction. In this paper, a sparsity-constrained regularization approach for deconvolution of the returned lidar waveform and restoration of the target cross-section is presented. Primal-dual interior point methods are exploited to solve the resulting nonlinear convex optimization problem. The optimal regularization parameter is determined based on the L-curve method, which provides high consistency in varied conditions. Quantitative evaluation and visual assessment of results show the superior performance of the proposed regularization approach in both removal of the effect of the system waveform and reconstruction of the target cross-section as compared to other prominent deconvolution approaches. This demonstrates the potential of the proposed approach for improving the accuracy of both range measurements and geophysical attribute retrieval. The feasibility and consistency of the presented approach in the processing of a variety of lidar data acquired under different system configurations is also highlighted.