Earth and Space Science and Engineering

During the past several years, point density covering topographic objects with airborne lidar (Light Detection And Ranging) technology has been greatly improved. This achievement provides an improved ability for reconstructing more complicated building roof structures; more specifically, those comprising various model primitives horizontally and/or vertically. However, the technology for automatically reconstructing such a complicated structure is thus far poorly understood and is currently based on employing a limited number of pre-specified building primitives. This paper addresses this limitation by introducing a new technique of modeling 3D building objects using a data-driven approach whereby densely collecting low-level modeling cues from lidar data are used in the modeling process. The core of the proposed method is to globally reconstruct geometric topology between adjacent linear features by adopting a BSP (Binary Space Partitioning) tree. The proposed algorithm consists of four steps: (a) detecting individual buildings from lidar data, (b) clustering laser points by height and planar similarity, (c) extracting rectilinear lines, and (d) planar partitioning and merging for the generation of polyhedral models. This paper demonstrates the efficacy of the algorithm for creating complex models of building rooftops in 3D space from airborne lidar data.

An automated method for boundary representation of building objects has been considered as a core processor for 3D city modelling. Since the reconstruction of generic building shape fundamentally depends on geometric features extracted from data sources, it suffers difficulties ...

This paper introduced a developed object-based extraction tool for delineating the building footprints from high-resolution satellite image. The foundation of the algorithm is a scale-space analysis based on level set transformation integrated with index, size, shape and spatial relation analysis. Building extraction tool is developed as a component of a newly developed tsunami early damage detection system. To be capable as a tool in such critical time after a catastrophe, the tool is implemented in a parallel manner for processing speed optimization. The extracted building footprint plays as one of the roughness factors in tsunami numerical modeling. The test on Quick-Bird image of Ban Nam Ken, Phanga, Thailand presents a good result in comparison with manually detected one. It infers that the tool is applicable in extraction of other object types from high-resolution satellite images. It is also recommended that benchmark should be carried out to report the algorithm performance based on which improvement can be made.

The co-registration process between light detection and ranging point clouds and photogrammetric digital surface models data is analyzed and a semi-automated solution is implemented. For a robust 3D geometric transformation between the two datasets in an urban environment, both planes and points are used. Initially, planes are chosen as the coregistration primitives. A region-growing algorithm based on a triangulated irregular network is implemented to extract planes from both datasets. Next, an automatic process for identifying and matching corresponding planes from the two datasets has been developed and implemented. The extracted planes are associated as plane pairs, initially by a matching process for buildings, followed by the plane matching algorithm within the building spatial window. Then two different geometric registration algorithms are used to obtain accurate transformation parameters between the two planar datasets. The 3D conformal transformation is applied to obtain the transformation parameters using the corresponding plane pairs. Following the mapping of one dataset into the coordinate system of the other, the mapped original point clouds are also used as another registration feature to complement the plane-based co-registration. In this latter step, the iterative closest point algorithm is applied, using the corresponding building point clouds to further refine the transformation solution. Experimental results together with their assessments are presented and discussed demonstrating the applicability of the proposed approach.

Wikification of GIS by the masses' is a phrase-term first coined by Kamel Boulos in 2005, two years earlier than Goodchild's term 'Volunteered Geographic Information'. Six years later (2005)(2006)(2007)(2008)(2009)(2010)(2011), OpenStreetMap and Google Earth (GE) are now full-fledged, crowdsourced 'Wikipedias of the Earth' par excellence, with millions of users contributing their own layers to GE, attaching photos, videos, notes and even 3-D (three dimensional) models to locations in GE. From using Twitter in participatory sensing and bicycle-mounted sensors in pervasive environmental sensing, to creating a 100,000-sensor geo-mashup using Semantic Web technology, to the 3-D visualisation of indoor and outdoor surveillance data in real-time and the development of next-generation, collaborative natural user interfaces that will power the spatially-enabled public health and emergency situation rooms of the future, where sensor data and citizen reports can be triaged and acted upon in real-time by distributed teams of professionals, this paper offers a comprehensive state-of-the-art review of the overlapping domains of the Sensor Web, citizen sensing and 'human-in-the-loop sensing' in the era of the Mobile and Social Web, and the roles these domains can play in environmental and public health surveillance and crisis/disaster informatics. We provide an in-depth review of the key issues and trends in these areas, the challenges faced when reasoning and making decisions with real-time crowdsourced data (such as issues of information overload, "noise", misinformation, bias and trust), the core technologies and Open Geospatial Consortium (OGC) standards involved (Sensor Web Enablement and Open GeoSMS), as well as a few outstanding project implementation examples from around the world.

For more than two decades, many efforts have been made to develop methods for extracting urban objects from data acquired by airborne sensors. In order to make the results of such algorithms more comparable, benchmarking data sets are of paramount importance. Such a data set, consisting of airborne image and laserscanner data, has been made available to the scientific community by ISPRS WGIII/4. Researchers were encouraged to submit their results of urban object detection and 3D building reconstruction, which were evaluated based on reference data. This paper presents the outcomes of the evaluation for building detection, tree detection, and 3D building reconstruction. The results achieved by different methods are compared and analysed to identify promising strategies for automatic urban object extraction from current airborne sensor data, but also common problems of state-of-the-art methods.

This study developed a new method to derive forest type information from large-footprint lidar data based on full waveform analysis. For this purpose, the raw waveform was decomposed into Gaussian components, and canopy return and ground return of the waveforms were separated. Two types of metrics hypothesized to have relationship with forest types were derived from the canopy return part of the waveform. The first type of metrics is quantile-based metrics reflecting the vertical distribution of canopy return energy, and the second type is statistical characteristics of the Gaussian components of canopy return part. Support Vector Machine classification was applied to different combinations of the metrics to find their relationship with different forest types. The results showed that the second type of metrics, indicating the canopy stratum characteristics, showed great promise in separating broad-leaved and needle-leaved forests with the accuracy ranging from 88.68 percent to 90.57 percent and Kappa statistic from 0.7406 to 0.7868.

A lidar filtering technique is used to differentiate on-terrain points and offterrain points from a cloud of 3D point data collected by a lidar system. A major issue of concern in this low-level filter is to design a methodology to ensure a continual adaptation to variations of terrain slopes and object scales. In this paper, a new lidar filtering technique which hierarchically fragments lidar data into piecewise planar terrain models is introduced. Once a number of hypothetical planar terrain models are generated to fit the terrain surface of the underlying area, the optimal terrain model to produce the minimum labelling errors is determined based on minimum description length (MDL) principles. This hypothesis-verification optimisation is achieved in a coarse-to-fine strategy by which the entire terrain surface is incrementally reconstructed by increasing the number of planar terrain models fitted. The proposed technique was successfully applied to a digital surface model provided within an OEEPE lidar trial, showing 0AE94% of Type I errors and 6AE75% of Type II errors compared to manually classified reference data.

Tree detection and reconstruction is of great interest in large-scale city modelling. In this paper, we present a marked point process model to detect single trees from airborne laser scanning (ALS) data. We consider single trees in ALS recovered canopy height model (CHM) as a realization of point process of circles. Unlike traditional marked point process, we sample the model in a constraint configuration space by making use of image process techniques. A Gibbs energy is defined on the model, containing a data term which judge the fitness of the model with respect to the data, and prior term which incorporate the prior knowledge of object layouts. We search the optimal configuration through a steepest gradient descent algorithm. The presented hybrid framework was test on three forest plots and experiments show the effectiveness of the proposed method.

The safety of powerline infrastructure significantly affects our everyday life and industrial activities. There are many factors and objects to threaten powerline safety, which include encroaching vegetation, tree healthiness, ambient temperature of the powerlines, structural faults of insulator and tower and so on. A timely and accurate monitoring of those key powerline features enables to prevent causing possible dangerous situation such as blackout. At present, most of utility firms heavily rely on men-centric powerline monitoring methods which are time consuming and very costly, and also, hazardous work. Recently, airborne LiDAR system was introduced as a cost effective data acquisition tool which enables to rapidly capture 3D powerline scene with up to about 30 points/m 2. This dramatically increased point density would provide a great possibility for achieving the automation of 3D reconstruction of powerline scene features which is an essential step for a machine-based powerline safety monitoring. Since it has been lately used for the powerline monitoring, not many automatic algorithms for reconstructing powerline have been introduced using LiDAR data. This paper introduces a Voxel-based Piece-wise Line Detector (VPLD) which automatically reconstructs 3D powerline models using airborne LiDAR data. The VPLD is developed based on well-known perceptual grouping framework which reconstructs a powerline by grouping similar features at different levels of information (i.e., points, segments and lines). A final reconstruction of single powerline models is accomplished by applying a non-linear adjustment for estimating catenary line parameters to a piece-wisely segmented voxel space. An evaluation of the proposed approach over a complicated powerline scene shows that the proposed method is promising.

The demands for automatic point cloud classification have dramatically increased with the widespread use of airborne LiDAR. Existing research has mainly concentrated on a few dominant objects such as terrain, buildings and vegetation. In addition to those key objects, this paper proposes a supervised classification method to identify other types of objects including power-lines and pylons from point clouds using a JointBoost classifier. The parameters for the learning model are estimated with various features computed based on the geometry and echo information of a LiDAR point cloud. In order to overcome the shortcomings stemming from the inclusion of bare ground data before classification, the proposed classifier directly distinguishes terrain using a feature step-off count. Feature selection is conducted using JointBoost to evaluate feature correlations thus improving both classification accuracy and operational efficiency. In this paper, the contextual constraints for objects extracted by graph-cut segmentation are used to optimize the initial classification results obtained by the JointBoost classifier. Our experimental results show that the step-off count significantly contributes to classification. Seventeen effective features are selected for the initial classification results using the JointBoost classifier. Our experiments indicate that the proposed features and method are effective for classification of airborne LiDAR data from complex scenarios.

This study presents a hybrid framework for single tree detection from airborne laser scanning (ALS) data by integrating low-level image processing techniques into a high-level probabilistic framework. The proposed approach modelled tree crowns in a forest plot as a configuration of circular objects. We took advantage of low-level image processing techniques to generate candidate configurations from the canopy height model (CHM): the treetop positions were sampled within the over-extracted local maxima via local maxima filtering, and the crown sizes were derived from marker-controlled watershed segmentation using corresponding treetops as markers. The configuration containing the best possible set of detected tree objects was estimated by a global optimization solver. To achieve this, we introduced a Gibbs energy, which contains a data term that judges the fitness of the objects with respect to the data, and a prior term that prevents severe overlapping between tree crowns on the configuration space. The energy was then embedded into a Markov Chain Monte Carlo (MCMC) dynamics coupled with a simulated annealing to find its global minimum. In this research, we also proposed a Monte Carlo-based sampling method for parameter estimation. We tested the method on a temperate mature coniferous forest in Ontario, Canada and also on simulated coniferous forest plots with different degrees of crown overlap. The experimental results showed the effectiveness of our proposed method, which was capable of reducing the commission errors produced by local maxima filtering, thus increasing the overall detection accuracy by approximately 10% on all of the datasets.

This study aims to introduce new methods for classifying key features (power lines, pylons, and buildings) comprising utility corridor scene using airborne LiDAR data and modelling power lines in 3D object space. The proposed approach starts from PL scene segmentation using Markov Random Field (MRF), which emphasizes on the roles of spatial context of linear and planar features as in a graphical model. The MRF classifier identifies power line features from linear features extracted from given corridor scenes. The non-power line objects are then investigated in a planar space to sub-classify them into building and non-building class. Based on the classification results, precise localization of individual pylons is conducted through investigating a prior knowledge of contextual relations between power line and pylon. Once the pylon localization is accomplished, a power line span is identified, within which power lines are modelled with catenary curve models in 3D. Once a local catenary curve model is established, this initial model progressively extends to capture entire power line points by adopting model hypothesis and verification. The model parameters are adjusted using a stochastic non-linear square method for producing 3D power line models. An evaluation of the proposed approach is performed over an urban PL corridor area that includes a complex PL scene.

Since the introduction of Airborne Laser Scanning (ALS) know as an alternative aerial-based data acquisition tool, the requirement of the 3D model reconstruction in both urban and power-line scenes has dramatically increased. Especially, electric utilities including power-line and tower are crucial infrastructures that require considerable resources to be monitored and managed effectively. For the establishment of the power-line scene inventory, its geospatial information such as positions and attributes of power-line networks should be accurately recorded. This paper presents a 3D classification method to classify power-line scene where a few structures including trees, transmission lines and pylons would be vertically overlapped. The research proposes two different scales of feature extractions from a volumetric space and its embedded points for taking advantages of full 3D analysis against conventional 2D pixel-based analysis. With targeted object instances including ground, vege...

The specific requirements of UAV-photogrammetry necessitate particular solutions for system development, which have mostly been ignored or not assessed adequately in recent studies. Accordingly, this paper presents the methodological and experimental aspects of correctly implementing a UAV-photogrammetry system. The hardware of the system consists of an electric-powered helicopter, a high-resolution digital camera and an inertial navigation system. The software of the system includes the in-house programs specifically designed for camera calibration, platform calibration, system integration, on-board data acquisition, flight planning and on-the-job self-calibration. The detailed features of the system are discussed, and solutions are proposed in order to enhance the system and its photogrammetric outputs. The developed system is extensively tested for precise modeling of the challenging environment of an open-pit gravel mine. The accuracy of the results is evaluated under various ma...

With rapid urbanization, highly accurate and semantically rich virtualization of building assets in 3D become more critical for supporting various applications, including urban planning, emergency response and location-based services. Many research efforts have been conducted to automatically reconstruct building models at city-scale from remotely sensed data. However, developing a fully-automated photogrammetric computer vision system enabling the massive generation of highly accurate building models still remains a challenging task. One the most challenging task for 3D building model reconstruction is to regularize the noises introduced in the boundary of building object retrieved from a raw data with lack of knowledge on its true shape. This paper proposes a data-driven modeling approach to reconstruct 3D rooftop models at city-scale from airborne laser scanning (ALS) data. The focus of the proposed method is to implicitly derive the shape regularity of 3D building rooftops from ...

This paper proposes a new algorithm to generalize noisy polylines comprising a rooftop model by maximizing a shape regularity (orthogonality, symmetricity and directional simplications). The nature of remotely sensed data including airborne LiDAR often produce errors in localizing salient features (corners, lines and planes) due to weak contrast, occlusions, shadows and object complexity. A generalization or regularization process is well known algorithm for eliminating erroneous vertices while preserving significant information on rooftop shapes. Most of existing regularization methods achieves this goal base on a local process such as if-then rules due to lacking global objective functions or mainly focusing on minimising residuals between boundary observations and models. In this study, we implicitly derive rules to generate local hypothetical models. Those hypothesized models produce possible drawings of regular patterns that given rooftop vectors can possibly generate by combining global and local analysis of line directions and their connections. A final optimal model is globally selected through a gradient descent optimization. A BSP (Binary Tree Partitioning)-tree was used to produce initial rooftop vectors using ISPRS WGIII/4's benchmarking test sites in Veihngen. The proposed regularization algorithm was applied to reduce modelling errors produced by BSP-tree. An evaluation demonstrates the proposed algorithm is promising for updating of building database.

In this work, the co-registration steps between LiDAR and photogrammetric DSM 3Ddata are analyzed and a solution based on automated plane matching is proposed and implemented. For a robust 3D geometric transformation both planes and points are used. Initially planes are chosen as the co-registration primitives. To confine the search space for the plane matching a sequential automatic building matching is performed first. For matching buildings from the LiDAR and the photogrammetric data, a similarity objective function is formed based on the roof height difference (RHD), the 3D histogram of the building attributes, and the building boundary area of a building. A region growing algorithm based on a Triangulated Irregular Network (TIN) is implemented to extract planes from both datasets. Next, an automatic successive process for identifying and matching corresponding planes from the two datasets has been developed and implemented. It is based on the building boundary region and determines plane pairs through a robust matching process thus eliminating outlier pairs. The selected correct plane pairs are the input data for the geometric transformation process. The 3D conformal transformation method in conjunction with the attitude quaternion is applied to obtain the transformation parameters using the normal vectors of the corresponding plane pairs. Following the mapping of one dataset onto the coordinate system of the other, the Iterative Closest Point (ICP) algorithm is then applied, using the corresponding building point clouds to further refine the transformation solution. The results indicate that the combination of planes and points improve the co-registration outcomes.

The increasing use of electrical energy has yielded more necessities of electric utilities including transmission lines and electric pylons which require a real-time risk monitoring to prevent massive economical damages. Recently, Airborne Laser Scanning (ALS) has become one of primary data acquisition tool for corridor mapping due to its ability of direct 3D measurements. In particular, for power-line risk management, a rapid and accurate classification of power-line objects is an extremely important task. We propose a 3D classification method combining results obtained from multiple classifier trained with different features. As a base classifier, we employ Random Forests (RF) which is a composite descriptors consisting of a number of decision trees populated through learning with bootstrapping samples. Two different sets of features are investigated that are extracted in a point domain and a feature (i.e., line & polygon) domain. RANSAC and Minimum Description Length (MDL) are applied to create lines and a polygon in each volumetric pixel (voxel) for the line & polygon features. Two RFs are trained from the two groups of features uncorrelated by Principle Component Analysis (PCA), which results are combined for final classification. The experiment with two real datasets demonstrates that the proposed classification method shows 10% improvements in classification accuracy compared to a single classifier.