Millions Of Points Research Articles

Computing 2D Constrained Delaunay Triangulation Using the GPU

We propose the first graphics processing unit (GPU) solution to compute the 2D constrained Delaunay triangulation (CDT) of a planar straight line graph (PSLG) consisting of points and edges. There are many existing CPU algorithms to solve the CDT problem in computational geometry, yet there has been no prior approach to solve this problem efficiently using the parallel computing power of the GPU. For the special case of the CDT problem where the PSLG consists of just points, which is simply the normal Delaunay triangulation (DT) problem, a hybrid approach using the GPU together with the CPU to partially speed up the computation has already been presented in the literature. Our work, on the other hand, accelerates the entire computation on the GPU. Our implementation using the CUDA programming model on NVIDIA GPUs is numerically robust, and runs up to an order of magnitude faster than the best sequential implementations on the CPU. This result is reflected in our experiment with both randomly generated PSLGs and real-world GIS data having millions of points and edges.

Computed tomography in quality control: chances and challenges

Since the first applications of computed tomography for medical diagnostics in the 1970s, the technology has been rapidly adopted for non-destructive testing of casted workpieces or complex assemblies. In 2005, the first coordinate measuring machine with computed tomography sensor has been presented for manufacturing metrology and quality control. Today, there exist numerous different manufacturers as the benefits for metrology are convincing. Computed tomography offers the possibility to acquire a measurement object holistically, that is, the whole volume of the object, not only the surface, with a very high point density of typically several millions of points. Current research and development activities are focused on several different aspects of computed tomography. First, the machine components are analysed and improved with respect to application in metrology. Larger detectors with better resolution or X-ray tubes with smaller focal spots may help in improving the machine’s accuracy. Second, algorithms for reconstruction and artefact reduction are developed, especially for metrological purpose. As technical measurement objects feature a greater variety of material properties than in medical diagnostics, different strategies have to be investigated. The most crucial point is that not only the image quality of the radiographies or the reconstructed data is important but also the measurement accuracy plays the decisive role. Finding the interrelation between image quality and measurement accuracy is an important research topic. Third, from a metrological point view, the determination of measurement accuracy and measurement uncertainty for computed tomography measurements is an important task. Methods for experimental and simulation-based uncertainty determination are developed at the time being.

Monitoring Of Complex Structure For Structural Control Using Terrestrial Laser Scanning (Tls) And Photogrammetry

The demand for high-definition surveys within cultural heritage-related projects represents one of the main factors that promoted the use of laser scanning technology and photogrammetry. By measuring millions of points within relatively short time periods, terrestrial laser scanners (TLS) allows to researchers to derive complete and very detailed three-dimensional (3D) models of real objects from acquired point clouds. These features drew in recent years the interest of surveyors, engineers, architects, and archaeologists towards the laser scanning technique as an invaluable surveying tool for 3D modeling of sites and artifacts of cultural heritage. A wide variety of objects, such as small pieces of pottery, statues, buildings, and large areas of archaeological sites, have been scanned and modeled for various purposes like preservation, reconstruction, study, and museum exhibitions. However, the use of TLS systems for stability control is still a research field not much investigated. In the view of in-depth investigation on this topic, a 3-years project has been established to evaluate the use of multiple surveying techniques for the stability control of a complex historical structure. To this aim, TLS, total station (TS) and photogrammetry are being employed for stability control monitoring with finite element model (FEM) analysis applied to an historical building, Teatro Olimpico (Olympic Theatre), in Vicenza, Italy. The main goal of this work is to analyze and verify the stability over time of this kind of structure by applying FEM analysis to a highly detailed 3D model of the theater. To date, three consecutive surveys of the theater have been carried out with consumer digital-reflex camera Nikon D200, a Leica Laser Scanner (HDS 3000) and a Leica Total Station (TCR 705). The first survey comprised approximately 250 pictures to derive a global and complete 3D model of theater with an inexpensive measuring and modeling photogrammetry software (Photomodeler). In the second study, the historical structure was fully surveyed with a TLS, the Leica HDS 3000, to produce a complete 3D model. A set of scans of complex elements, such as the wood trompe l'oeil onstage scenery, statues, and fine trim, were acquired in this stage. This article presents the results from the repeated surveys and highlights the issues and difficulties related to the laser scanning and photogrammetry of an unusual and complex geometry such as the one provided by the Olympic Theater in Vicenza.

Reconstruction of 3D Natural Field Scenes Based on SLAM Robot

Abstract Mobile robots have already been widely applied in agriculture fields. In order to perform complex tasks such as localization, mapping and path planning, mobile robots need to understand their environment. Therefore, the perception ability is important to robots. In the past twenty years, Simultaneous Localization and Mapping has been a popular algorithm to solve localization and mapping problems. In agricultural applications, robots need to overcome bumpy surfaces and should avoid inducing destruction to plant growths. Hence, the assumptions of robot poses need to be represented by six degree of freedom, not just flat surface movements. In order to estimate the 6D poses, 3D information was acquired by our custom-made of 3D laser scanner. The ICP algorithm then merged different scenes together to reconstruct the whole environment. However, errors are accumulated in the ICP algorithm, so a global optimization algorithm was applied to deal with this accumulated error. We had successfully reconstructed two large scale environments that contain millions of points. The method we proposed here has a RMSE of 55.94 centimeters. It is believed that this system has different applications such as agriculture field investigation, virtual reality and geographic information system.

Numerical solutions of the γ-index in two and three dimensions

The γ-index is used routinely to establish correspondence between two dose distributions. The definition of the γ-index can be written with a single equation but solving this equation at millions of points is computationally expensive, especially in three dimensions. Our goal is to extend the vector–equation method in Bakai et al (2003 Phys. Med. Biol.48 3543–53) to higher order for better accuracy and, as important, to determine the magnitude of accuracy in a higher order solution. We construct a numerical framework for calculating the γ-index in two and three dimensions and present an efficient method for calculating the γ-index with zeroth-, first- and second-order methods using tricubic spline interpolation. For an intensity-modulated radiation therapy example with 1.78 × 106 voxels, the zeroth-order, first-order, first-order iterations and semi-second-order methods calculate the three-dimensional γ-index in 1.5, 4.7, 34.7 and 35.6 s with 36.7%, 1.1%, 0.2% and 0.8% accuracy, respectively. The accuracy of linear interpolation with this example is 1.0%. We present efficient numerical methods for calculating the three-dimensional γ-index with tricubic spline interpolation. The first-order method with iterations is the most accurate and fastest choice of the numerical methods if the dose distributions may have large second-order gradients. Furthermore, the difference between iterations can be used to determine the accuracy of the method.

Appropriate formulation of the objective function for the history matching of seismic attributes

The purpose of history matching is to find one or several reservoir models which can reproduce as best as possible all the available data. The available data are traditionally some production data, but today seismic data are often integrated in the history matching process. The way of measuring the misfit between real data and simulated responses has a significant impact on the optimization process and hence on the final optimal model obtained. The classical formulation of the misfit is the least square one, which was used with success for production data. This formulation was naturally extended for seismic data. However, it yields an objective function term which is difficult to reduce. Indeed, seismic data are different from production data since they are defined by millions of points and are generally very noisy. When matching seismic data, the goal is then to capture the main features. As a result, computing a point to point error is not adapted and the resulting objective function is not representative of the quality expected for the match. We propose in this paper to define a more appropriate formulation. The idea is to use some image analysis tools to define a formulation focusing on the main features of seismic images. More precisely, it is based on image segmentation and on a modified Hausdorff metric. We illustrate the success of this formulation on a simple history matching case.

High-performance visual analytics of terrestrial light detection and ranging data on large display wall

A typical LIDAR (light detection and ranging) scan contains hundreds of millions of points. As such, the visualization of LIDAR point clouds poses a significant challenge in data analysis. We propose to visualize and process LIDAR point clouds on a large display wall with an array of monitors. This provides researchers with a high-resolution display environment for looking at and studying large data sets. High-resolution large displays offer both global perspectives and local details of point clouds, which is essential in the process of data exploration. The ability to explore, conceptualize, and correlate spatial and temporal changes of topographical records is required for the development of new analytical models that capture the mechanisms contributing toward cliff erosion. Large displays driven by high-performance parallel visualization cluster allow researchers to fully interact with LIDAR point clouds of slopes in Houshanyue mountain and cliff failures observed in Solana Beach in California. In our study, cases studies of visualization based approaches were conducted using large displays in digital immersive environments. Visual analytics techniques such as delineation, segmentation, and classification of features, change detection, and annotation were used to perform erosion assessment. The results showed that the researchers can observe the temporal change of a failure mass effectively in high-resolution large display environments.

Creating Large-Scale City Models from 3D-Point Clouds: A Robust Approach with Hybrid Representation

We present a novel and robust method for modeling cities from 3D-point data. Our algorithm provides a more complete description than existing approaches by reconstructing simultaneously buildings, trees and topologically complex grounds. A major contribution of our work is the original way of modeling buildings which guarantees a high generalization level while having semantized and compact representations. Geometric 3D-primitives such as planes, cylinders, spheres or cones describe regular roof sections, and are combined with mesh-patches that represent irregular roof components. The various urban components interact through a non-convex energy minimization problem in which they are propagated under arrangement constraints over a planimetric map. Our approach is experimentally validated on complex buildings and large urban scenes of millions of points, and is compared to state-of-the-art methods.

Efficient rate-distortion compression of dynamic point cloud for grid-pattern-based 3D scanning systems

Recently 3D scanning systems are capable of modeling entire dense shapes that evolve over time with a single scan (a.k.a. one-shot scan). In particular, structured-light-based systems have emerged as one-shot shape reconstruction methods that project a static grid pattern onto the object surface. This pattern allows the scanning of moving objects while still maintaining dense reconstruction. One-shot scanning systems are then capable of producing 3D+t (a.k.a. 4D) spatio-temporal models with millions of points. As a consequence, effective 4D geometry compression schemes are required to face the need to store or transmit the huge amount of data, in addition to classical static 3D data. In this paper, we propose a 4D spatiotemporal rate-distortion (RD) optimized point cloud encoder via a curve-based representation of the point cloud, particularly well-suited for one-shot scanning systems. The object surface is naturally sampled in a series of curves, due to the grid pattern. This motivates our choice to leverage a curve-based representation to remove the spatial and temporal correlation of the sampled point along the scanning directions through a competitive-based predictive encoder that includes different spatio-temporal prediction modes through an RD cost computation control. Experimental results show the significant gain obtained with the proposed method.

Fast and automatic image-based registration of TLS data

The fast and automatic registration of laser scanner data is of great interest in photogrammetric research. Recent developments show that for registration purposes characteristic 3D points can be extracted from the measured laser data, where the data are also represented as image. In this paper, radiometric and geometric information derived from TLS data are utilized for estimating the transformation parameters between two unregistered point clouds. After the extraction of characteristic 2D points based on SIFT features, these points are projected into 3D space by using interpolated range information. From these 3D conjugate points and their corresponding 2D projections onto a virtual plane 3D-to-2D correspondences are established. The fast, accurate and robust RANSAC-based registration scheme including the EPnP algorithm provides a framework to estimate the coarse transformation parameters from these 3D-to-2D correspondences. The coarse estimates are further refined by a single step outlier removal to gain a higher accuracy by introducing additional geometric constraints. These new constraints are based on 3D-to-3D correspondences which are much stronger than the 3D-to-2D correspondences alone. It will be shown that the presented approach is successfully applied to a benchmarked data set with millions of points resulting in a fast and accurate estimation of the transformation parameters with a processing speed of several seconds on a standard PC and an accuracy in the low centimeter range.

Regulating the disposal of cigarette butts as toxic hazardous waste

The trillions of cigarette butts generated each year throughout the world pose a significant challenge for disposal regulations, primarily because there are millions of points of disposal, along with the...

Efficient Mean‐shift Clustering Using Gaussian KD‐Tree

AbstractMean shift is a popular approach for data clustering, however, the high computational complexity of the mean shift procedure limits its practical applications in high dimensional and large data set clustering. In this paper, we propose an efficient method that allows mean shift clustering performed on large data set containing tens of millions of points at interactive rate. The key in our method is a new scheme for approximating mean shift procedure using a greatly reduced feature space. This reduced feature space is adaptive clustering of the original data set, and is generated by applying adaptive KD‐tree in a high‐dimensional affinity space. The proposed method significantly reduces the computational cost while obtaining almost the same clustering results as the standard mean shift procedure. We present several kinds of data clustering applications to illustrate the efficiency of the proposed method, including image and video segmentation, static geometry model and time‐varying sequences segmentation.

Determining leaf area index and leafy tree roughness using terrestrial laser scanning

Vegetation roughness, and more specifically forest roughness, is a necessary component in better defining flood dynamics both in the sense of changes in river catchment characteristics and the dynamics of forest changes and management. Extracting roughness parameters from riparian forests can be a complicated process involving different components for different required scales and flow depths. For flow depths that enter a forest canopy, roughness at both the woody branch and foliage level is necessary. This study attempts to extract roughness for this leafy component using a relatively new remote sensing technique in the form of terrestrial laser scanning. Terrestrial laser scanning is used in this study due to its ability to obtain millions of points within relatively small forest stands. This form of lidar can be used to determine the gaps present in foliaged canopies in order to determine the leaf area index. The leaf area index can then be directly input into resistance equations to determine the flow resistance at different flow depths. Leaf area indices created using ground scanning are compared in this study to indices calculated using simple regression equations. The dominant riparian forests investigated in this study are planted and natural poplar forests over a lowland section of the Garonne River in Southern France. Final foliage roughness values were added to woody branch roughness from a previous study, resulting in total planted riparian forest roughness values of around Manning's n = 0.170–0.195 and around n = 0.245–330 for in‐canopy flow of 6 and 8 m, respectively, and around n = 0.590 and around n = 0.750 for a natural forest stand at the same flow depths.

Visual analysis of flow features using information theory.

Over the past decades, scientific visualization has helped tremendously to easily generate meaningful representations of complicated data sets. However, with data correlated over many dimensions and millions of points, only few of the standard techniques are directly applicable. Unsteady multifield visualizations require effective reduction of the data to be displayed. From a huge amount of information, scientists must be able to extract the most informative parts. ?-machines, a concept based on information theory, can handle this task. They're a finitestate machine representation of a system's dynamics, which can be represented as a directed graph (see Figure 1). The nodes encode the local dynamics given as a spatiotemporal stochastic pattern, and the edges indicate the flow's evolution. ?-machines consist of causal states and transitions between them. Several enhancements to the fundamental ?-machine representation can help users identify interesting time intervals, analyze the evolution of unusual local dynamics, and track features over time.

Identification of preconditions for an emerging mobile LBS market

Location information has become a very important horizontal service component for a wide range of vertical mobile applications. This enabled multiple innovative mobile location-based services (LBS) that range from conventional mapping and navigation services featuring millions of points of interest (POI) to location-aware content delivery and mobile social networking. Providing a whole new world of opportunities for device and application vendors, network operators, content providers and prospective consumers, mobile LBS have potential to become one of the core businesses in today's mobile arena. However, the future of this market is not fully transparent due to several uncertainties associated with a mixture of technologies, service demand, consumer attitude, privacy aspects of using LBS and a complex value network. In this study, we identify and analyse a set of preconditions for an emerging mobile LBS market in its present evolution phase. The history of LBS motivates us to take a closer look at this emerging market and try to understand what the main reasons for LBS initial failure have been and what is still missing in this market today to emerge successfully. To focus on the most relevant issues, we construct our own research framework containing three main dimensions: technological advance and standardisation, value network and value proposition. Within these dimensions, we compare the early state of mobile LBS with the present situation and analyse the key factors presently crucial for LBS proliferation.

On optimal service selection in Service Oriented Architectures

Service Oriented Architectures enable a multitude of service providers to provide loosely coupled and interoperable services at different Quality of Service and cost levels. This paper considers business processes composed of activities that are supported by service providers. The structure of a business process may be expressed by languages such as BPEL and allows for constructs such as sequence, switch, while, flow, and pick. This paper considers the problem of finding the set of service providers that minimizes the total execution time of the business process subject to cost and execution time constraints. The problem is clearly NP-hard. However, the paper presents an optimized algorithm that finds the optimal solution without having to explore the entire solution space. This algorithm can be used to find the optimal solution in problems of moderate size. A heuristic solution is also presented. Thorough experimental studies, based on random business processes, demonstrate that the heuristic algorithm was able to produce service provider allocations that result in execution times that are only a few percentage points (less than 2.5%) worse than the allocations obtained by the optimal algorithm while examining a tiny fraction of the solution space (tens of points versus millions of points).

Extensions of Parallel Coordinates for Interactive Exploration of Large Multi-Timepoint Data Sets

Parallel coordinate plots (PCPs) are commonly used in information visualization to provide insight into multi-variate data. These plots help to spot correlations between variables. PCPs have been successfully applied to unstructured datasets up to a few millions of points. In this paper, we present techniques to enhance the usability of PCPs for the exploration of large, multi-timepoint volumetric data sets, containing tens of millions of points per timestep. The main difficulties that arise when applying PCPs to large numbers of data points are visual clutter and slow performance, making interactive exploration infeasible. Moreover, the spatial context of the volumetric data is usually lost. We describe techniques for preprocessing using data quantization and compression, and for fast GPU-based rendering of PCPs using joint density distributions for each pair of consecutive variables, resulting in a smooth, continuous visualization. Also, fast brushing techniques are proposed for interactive data selection in multiple linked views, including a 3D spatial volume view. These techniques have been successfully applied to three large data sets: Hurricane Isabel (Vis'04 contest), the ionization front instability data set (Vis'08 design contest), and data from a large-eddy simulation of cumulus clouds. With these data, we show how PCPs can be extended to successfully visualize and interactively explore multi-timepoint volumetric datasets with an order of magnitude more data points.

Self-similarity based compression of point set surfaces with application to ray tracing

Many real-world, scanned surfaces contain repetitive structures, like bumps, ridges, creases, and so on. We present a compression technique that exploits self-similarity within a point-sampled surface. Our method replaces similar surface patches with an instance of a representative patch. We use a concise shape descriptor to identify and cluster similar patches. Decoding is achieved through simple instancing of the representative patches. Encoding is efficient, and can be applied to large data sets consisting of millions of points. Moreover, our technique offers random access to the compressed data, making it applicable to ray tracing, and easily allows for storing additional point attributes, like normals.

A numerical strategy for modelling rotating stall in core compressors

AbstractThe paper will focus on one specific core‐compressor instability, rotating stall, because of the pressing industrial need to improve current design methods. The determination of the blade response during rotating stall is a difficult problem for which there is no reliable procedure. During rotating stall, the blades encounter the stall cells and the excitation depends on the number, size, exact shape and rotational speed of these cells. The long‐term aim is to minimize the forced response due to rotating stall excitation by avoiding potential matches between the vibration modes and the rotating stall pattern characteristics. Accurate numerical simulations of core‐compressor rotating stall phenomena require the modelling of a large number of bladerows using grids containing several tens of millions of points. The time‐accurate unsteady‐flow computations may need to be run for several engine revolutions for rotating stall to get initiated and many more before it is fully developed. The difficulty in rotating stall initiation arises from a lack of representation of the triggering disturbances which are inherently present in aeroengines. Since the numerical model represents a symmetric assembly, the only random mechanism for rotating stall initiation is provided by numerical round‐off errors. In this work, rotating stall is initiated by introducing a small amount of geometric mistuning to the rotor blades. Another major obstacle in modelling flows near stall is the specification of appropriate upstream and downstream boundary conditions. Obtaining reliable boundary conditions for such flows can be very difficult. In the present study, the low‐pressure compression (LPC) domain is placed upstream of the core compressor. With such an approach, only far field atmospheric boundary conditions are specified which are obtained from aircraft speed and altitude. A chocked variable‐area nozzle, placed after the last compressor bladerow in the model, is used to impose boundary conditions downstream. Such an approach is representative of modelling an engine.Using a 3D viscous time‐accurate flow representation, the front bladerows of a core compressor were modelled in a whole‐annulus fashion whereas the rest of bladerows are modelled in a single‐passage fashion. The rotating stall behaviour at two different compressor operating points was studied by considering two different variable‐vane scheduling conditions for which experimental data were available. Using a model with nine whole‐assembly models, the unsteady‐flow calculations were conducted on 32‐CPUs of a parallel cluster, typical run times being around 3–4 weeks for a grid with about 60 million points. The simulations were conducted over several engine rotations. As observed on the actual development engine, there was no rotating stall for the first scheduling condition while mal‐scheduling of the stator vanes created a 12‐band rotating stall which excited the 1st flap mode. Copyright © 2006 John Wiley & Sons, Ltd.

Utilizing airborne scanning laser (LiDAR) to improve the estimation of Australian forest structure and biomass

Utilizing airborne scanning laser (LiDAR) to improve the estimation of Australian forest structure and biomass. Alex Lee. School of Resources, Environment and Society, Australian National University, Canberra, ACT 0200, Australia. Email: alex.lee@anu.edu.au. Key words: airborne LiDAR, biomass, condition, disturbance, forest structure, stand growth stage. The primary aim of the research is to develop algorithms to accurately quantify forest structure and above ground biomass using airborne scanning lasers (LiDAR), to assist regional and national vegetation monitoring initiatives. The utility of LiDAR is being examined across a range of scales; from within plot structural variability to broad structural types across environmental gradients. The research used LiDAR (1125 ha in central Queensland, and 60 000 ha in Broken and Ovens catchments of North-east Victoria) with field and other data. Northeast Victoria is where a National Forest Inventory pilot project, undertaken in mid-2003, is evaluating the possible implementation of a Continental Forest Monitoring Framework (BRS 2006). Three dimensional tree modelling combined with empirical relationships from tree and plot data are used to measure forest structure and aboveground biomass. In Queensland, multiple regression analyses utilized six canopy heights from overstory and understory with percentage crown cover to produce plot biomass estimates equivalent to field measurements (r2 = 0.92, SE = 12 Mg ha−1). Regional modification of the biomass function was required after testing at the Victorian site, where foliage cover, elevation, and five canopy heights were found to be optimal for biomass estimation in the taller, multilayered higher elevation forests. The research has concluded that LiDAR data can provide information just as detailed and possibly more accurately than field measurements for many required forest attributes. When utilized within sampling schemes much larger areas can be accurately reported on (Tickle et al. 2006). Information obtained from 3D tree and stand modelling using LiDAR data can be used to train satellite sensor (optical or radar) data to better report forest information at regional and national scales (Lucas et al. 2006). Current LiDAR use in Australia has primarily focused on accurately retrieving basic inventory attributes (Lovell et al. 2003), with little attention to vegetation condition assessments. This is mainly because of cost (data supply for this research was $53 ha−1 (Queensland) and $1.50 ha−1 (Victoria)), issues with processing large volumes of data (many millions of points), combined with condition attributes that are poorly defined or unsuitable for remote sensing. As condition is becoming increasingly important, the CSIRO and Victorian governments are investigating linking condition assessments (Habitat Hectares) with both ground and airborne LiDAR measures. LiDAR has advantages over traditional optical data because it can sample both the canopy (upper and lower strata) and ground, however, the representativeness depends on vegetation density and LiDAR collection parameters (e.g. beam size, scanning rate). Investigations of LiDAR apparent vertical foliage profiles show promise for the assessment of stand growth stage and understory recovery since disturbance (Parker & Russ 2004). As these are useful for forest condition assessment, they have been initially assessed both in Queensland (Lee et al. 2004) and Victoria. Here three relatively mature peppermint forest (Eucalyptus radiata Sieber ex DC) plots indicated a progression of understory top height since last recorded fire, from no understory after recent fire, to 10 m understorey after 12 years, to multiple strata (at 10 m and 20 m) after 64 years without fire.