Bernhard Höfle

Object-Based Point Cloud Analysis of Full-Waveform Airborne Laser Scanning Data for Urban Vegetation Classification.

Airborne laser scanning (ALS) is a remote sensing technique well-suited for 3D vegetation mapping and structure characterization because the emitted laser pulses are able to penetrate small gaps in the vegetation canopy. The backscattered echoes from the foliage, woody vegetation, the terrain, and other objects are detected, leading to a cloud of points. Higher echo densities (>20 echoes/m2) and additional classification variables from full-waveform (FWF) ALS data, namely echo amplitude, echo width and information on multiple echoes from one shot, offer new possibilities in classifying the ALS point cloud. Currently FWF sensor information is hardly used for classification purposes. This contribution presents an object-based point cloud analysis (OBPA) approach, combining segmentation and classification of the 3D FWF ALS points designed to detect tall vegetation in urban environments. The definition tall vegetation includes trees and shrubs, but excludes grassland and herbage. In the applied procedure FWF ALS echoes are segmented by a seeded region growing procedure. All echoes sorted descending by their surface roughness are used as seed points. Segments are grown based on echo width homogeneity. Next, segment statistics (mean, standard deviation, and coefficient of variation) are calculated by aggregating echo features such as amplitude and surface roughness. For classification a rule base is derived automatically from a training area using a statistical classification tree. To demonstrate our method we present data of three sites with around 500,000 echoes each. The accuracy of the classified vegetation segments is evaluated for two independent validation sites. In a point-wise error assessment, where the classification is compared with manually classified 3D points, completeness and correctness better than 90% are reached for the validation sites. In comparison to many other algorithms the proposed 3D point classification works on the original measurements directly, i.e. the acquired points. Gridding of the data is not necessary, a process which is inherently coupled to loss of data and precision. The 3D properties provide especially a good separability of buildings and terrain points respectively, if they are occluded by vegetation.

Automatic Roof Plane Detection and Analysis in Airborne Lidar Point Clouds for Solar Potential Assessment

A relative height threshold is defined to separate potential roof points from the point cloud, followed by a segmentation of these points into homogeneous areas fulfilling the defined constraints of roof planes. The normal vector of each laser point is an excellent feature to decompose the point cloud into segments describing planar patches. An object-based error assessment is performed to determine the accuracy of the presented classification. It results in 94.4% completeness and 88.4% correctness. Once all roof planes are detected in the 3D point cloud, solar potential analysis is performed for each point. Shadowing effects of nearby objects are taken into account by calculating the horizon of each point within the point cloud. Effects of cloud cover are also considered by using data from a nearby meteorological station. As a result the annual sum of the direct and diffuse radiation for each roof plane is derived. The presented method uses the full 3D information for both feature extraction and solar potential analysis, which offers a number of new applications in fields where natural processes are influenced by the incoming solar radiation (e.g., evapotranspiration, distribution of permafrost). The presented method detected fully automatically a subset of 809 out of 1,071 roof planes where the arithmetic mean of the annual incoming solar radiation is more than 700 kWh/m2.

Ontology-Based Classification of Building Types Detected from Airborne Laser Scanning Data

Abstract: Accurate information on urban building types plays a crucial role for urban development, planning, and management. In this paper, we apply Object-Based Image Analysis (OBIA) methods to extract buildings from Airborne Laser Scanner (ALS) data and investigate the possibility of classifying detected buildings into ―Residential/Small Buildings‖, ―Apartment Buildings‖, and ―Industrial and Factory Building‖ classes by means of domain ontology and machine learning techniques. The buildings objects are classified using exclusively the information computed from the ALS data. To select the relevant features for predicting the classes of interest, the Random Forest classifier has been applied. The ontology-based classification yielded convincing results for the ―Residential/Small Buildings‖ class (F-Measure 97.7%), whereas the ―Apartment Buildings‖ and ―Industrial and Factory Buildings‖ classes achieved less accurate results (F-Measure 60% and 51%, respectively).

Radiometric Correction of Terrestrial LiDAR Point Cloud Data for Individual Maize Plant Detection

Detailed geoinformation on in-field variations of plant properties (e.g., density, height) is required in precision agriculture and serves as a valuable input for plant growth models and crop management strategies. This letter presents a novel workflow for object-based point cloud analysis for individual maize plant mapping, using radiometric and geometric features of terrestrial laser scanning. The performed radiometric correction achieves a reduction of amplitude variation of homogeneous areas to 1/3 of the original variation and offers a distinct separability of the target class maize plant from soil. The developed procedure, including 3-D point cloud filtering and segmentation, is able to reliably detect single plants with a completeness and correctness . Experiments on reduced point densities show stability of detection rates above 100 points per 0.01 m2. The results indicate that the developed workflow will lead to even higher detection accuracy with LiDAR point clouds captured by mobile platforms, with less occlusion effects and more homogeneous point density.

Calibration of full-waveform airborne laser scanning data for object classification

Small-footprint airborne laser scanners with waveform-digitizing capabilities are becoming increasingly available. Waveform-digitizing laser scanners seize the physical measurement process in its entire complexity. This leads the way to the possibility of deriving the backscatter cross section which is a measure of the electromagnetic energy intercepted and reradiated by objects. The cross section can be obtained by firstly decomposing the echo waveform in several distinct echoes, whereas for each echo its range, amplitude and width are known. Then the radar equation can be used for calibrating the waveform measurements using external reference targets with known backscatter cross sections. The final outcome is a 3D point cloud where each point represents one scatterer with a given cross section and echo width. Using these physical attributes and various geometric criteria the point cloud can be segmented or classified. In this paper this procedure is demonstrated based on waveform measurements acquired by the RIEGL LMS-Q560 sensor. The cross section of the homogenous reference targets is estimated with a RIEGL reflectometer and Spectralon® targets.

Extraction of Vertical Walls from Mobile Laser Scanning Data for Solar Potential Assessment

In recent years there has been an increasing demand among home owners for cost effective sustainable energy production such as solar energy to provide heating and electricity. A lot of research has focused on the assessment of the incoming solar radiation on roof planes acquired by, e.g., Airborne Laser Scanning (ALS). However, solar panels can also be mounted on building facades in order to increase renewable energy supply. Due to limited reflections of points from vertical walls, ALS data is not suitable to perform solar potential assessment of vertical building facades. This paper focuses on a new method for automatic solar radiation modeling of facades acquired by Mobile Laser Scanning (MLS) and uses the full 3D information of the point cloud for both the extraction of vertical walls covered by the survey and solar potential analysis. Furthermore, a new method isintroduced determining the interior and exterior face, respectively, of each detected wall in order to calculate its slope and aspect angles that are of crucial importance for solar potential assessment. Shadowing effects of nearby objects are considered by computing the 3D horizon of each point of a facade segment within the 3D point cloud.

Estimation of aboveground biomass in alpine forests: a semi-empirical approach considering canopy transparency derived from airborne LiDAR data.

In this study, a semi-empirical model that was originally developed for stem volume estimation is used for aboveground biomass (AGB) estimation of a spruce dominated alpine forest. The reference AGB of the available sample plots is calculated from forest inventory data by means of biomass expansion factors. Furthermore, the semi-empirical model is extended by three different canopy transparency parameters derived from airborne LiDAR data. These parameters have not been considered for stem volume estimation until now and are introduced in order to investigate the behavior of the model concerning AGB estimation. The developed additional input parameters are based on the assumption that transparency of vegetation can bemeasured by determining the penetration of the laser beams through the canopy. These parameters are calculated for every single point within the 3D point cloud in order to consider the varying properties of the vegetation in an appropriate way. Exploratory Data Analysis (EDA) is performed to evaluate the influence of the additional LiDAR derived canopy transparency parameters for AGB estimation. The study is carried out in a 560 km2 alpine area in Austria, where reference forest inventory data and LiDAR data are available. The investigations show that the introduction of the canopy transparency parameters does not change the results significantly according to R2 (R2 = 0.70 to R2 = 0.71) in comparison to the results derived from, the semi-empirical model, which was originally developed for stem volume estimation.

Operational wide-area stem volume estimation based on airborne laser scanning and national forest inventory data

This paper evaluates the performance of a recently developed approach for wide-area stem volume estimations based on airborne laser scanning (ALS) and national forest inventory (NFI) data in the case where data recorded under operational conditions are used as input. This entails that neither ALS data nor NFI samples were collected and optimized for the current study. The approach was tested for the Austrian state of Vorarlberg, which covers an area of 2601 km2 and encloses about 970 km2 of forest land. ALS data with point densities varying between 1 and 4 points m−2 were acquired in the framework of a commercial state-wide terrain mapping project during several winter- and summer-flight campaigns. The stem volume model was calibrated with all NFI data available for Vorarlberg, whereas additional local forest inventory data were used for independent validation. Moreover, several relevant operational issues were addressed in this study, such as the determination of the optimum area used to calculate the reference laser metrics input to the model, the effect of gridding point cloud data to speed up processing, and the stratification of input data into coniferous and deciduous sample plots. Without tree species stratification and based on the 3D laser heights model, calibration provided a maximum R2 of 0.79 and a standard deviation (SD) of residuals derived from cross-validation of 107.4 m3 ha−1 (31.5%). Calibrating the model only with coniferous samples increased the achieved R2 to 0.81 and decreased SD to 104.8 m3 ha−1 (29.7%). As only eight NFI sample plots were available for deciduous forest a robust calibration of a separate model could not be obtained. Calibrating the model with a rasterized canopy height model (CHM) instead of using the 3D laser heights just led to a slight decrease in accuracy (R2 = 0.75, SD = 120.9 m3 ha−1 (35.5%) without forest-type stratification and R2 = 0.78 and SD = 117.2 m3 ha−1 (33.1%) for the coniferous stem volume model). Finally, the stem volume model calibrated with CHM data was adopted to generate a stem volume map of the entire State of Vorarlberg. Validation of this map with the additional local forest inventory data confirmed the accuracies (R2 = 0.75; SD = 135.6 m3 ha−1 (32.3%)) that were derived during calibration of the stem volume model based on the NFI data. The models and methods presented in this study are used operationally for forest and environment policy purposes and practical applications in Austria.

Area-wide roof plane segmentation in airborne LiDAR point clouds

Abstract Most algorithms performing segmentation of 3D point cloud data acquired by, e.g. Airborne Laser Scanning (ALS) systems are not suitable for large study areas because the huge amount of point cloud data cannot be processed in the computer’s main memory. In this study a new workflow for seamless automated roof plane detection from ALS data is presented and applied to a large study area. The design of the workflow allows area-wide segmentation of roof planes on common computer hardware but leaves the option open to be combined with distributed computing (e.g. cluster and grid environments). The workflow that is fully implemented in a Geographical Information System (GIS) uses the geometrical information of the 3D point cloud and involves four major steps: (i) The whole dataset is divided into several overlapping subareas, i.e. tiles. (ii) A raster based candidate region detection algorithm is performed for each tile that identifies potential areas containing buildings. (iii) The resulting building candidate regions of all tiles are merged and those areas overlapping one another from adjacent tiles are united to a single building area. (iv) Finally, three dimensional roof planes are extracted from the building candidate regions and each region is treated separately. The presented workflow reduces the data volume of the point cloud that has to be analyzed significantly and leads to the main advantage that seamless area-wide point cloud based segmentation can be performed without requiring a computationally intensive algorithm detecting and combining segments being part of several subareas (i.e. processing tiles). A reduction of 85% of the input data volume for point cloud segmentation in the presented study area could be achieved, which directly decreases computation time.

Roughness Mapping on Various Vertical Scales Based on Full-Waveform Airborne Laser Scanning Data

Abstract: Roughness is an important input parameter for modeling of natural hazards such as floods, rock falls and avalanches, where it is basically assumed that flow velocities decrease with increasing roughness. Seeing roughness as a multi-scale level concept ( i.e. , ranging from fine-scale soil characteristics to description of understory and lower tree layer) various roughness raster products were derived from the original full-waveform airborne laser scanning (FWF-ALS) point cloud using two different types of roughness parameters, the surface roughness ( SR ) and the terrain roughness ( TR ). For the calculation of the SR , ALS terrain points within a defined height range to the terrain surface are considered. For the parameterization of the SR, two approaches are investigated. In the first approach, a geometric description by calculating the standard deviation of plane fitting residuals of terrain points is used. In the second one, the potential of the derived echo widths are analyzed for the parameterization of