A methodology for evaluating range image segmentation algorithms is proposed. This involves (1) a common set of 40 laser finder images and structured light scanner that have manually specified ground truth (2) defined performance metrics instances correctly segmented, missed, noise regions, over- under-segmentation, accuracy the recovered geometry. tool used to objectively compare machine generated against truth. Four research groups contributed evaluate their own algorithm segmenting into...

Abstract. This paper presents a multi-stage approach toward the robust classification of travel-modes from GPS traces. Due to fact that traces are often composed more than one travel-mode, they segmented find sub-traces characterized as an individual travel-mode. is conducted by finding movement segments identifying stops. In first stage three main travel-mode classes identified: pedestrian, bicycle, and motorized vehicles; this achieved based on identified using speed, acceleration heading...

Currently the aspect graph is computed from theoretical standpoint of perfect resolution in object shape, viewpoint and projected image. This means that may include details an observer could never see practice. Introducing notion scale into framework provides a mechanism for selecting level detail "large enough" to merit explicit representation. effectively allows control over number nodes retained graph. paper introduces concept space graph, defines three different interpretations...

When reverse engineering a CAD model, it is necessary to integrate information from several views of an object into common reference frame. Given rough initial alignment, further pose refinement here uses improved version the interactive closes point algorithm. Incremental adjustments are computed simultaneously for all data sets, resulting in more globally optimal set transformations. Also, thresholds removing outlier correspondences not needed, as merging sets considered whole. Motion...

An algorithm for computing the aspect graph a class of curved-surface objects based on an exact parcellation 3-D viewpoint space is presented. The object considered solids revolution. A detailed analysis visual events this given, as well constructing graph. Numerical search techniques, geometric interpretation events, have been devised to determine those event surfaces that cannot be calculated directly. worst-case complexity number cells in space, and, hence, nodes graph, O(N/sup 4/), where...

This paper presents a procedure to extract parking statistics from 3D point cloud recorded with two 2D LiDAR sensors mounted on vehicle. Policy makers can use these reduce search traffic in cities by identifying characteristics and adjusting current rules policies. The extraction basically consists of an object segmentation classification step. For segmentation, region growing approach is used the ground surface separate distinct objects. classification, random forest classifier employed...

Several researchers have previously described approaches to 3-D object recognition which use nonlinear optimization control the matching of features a niodel found in an image. Recognition, this context, includes estimating parameters translation and orientation object. A major problem acknowledged by previous is how efficiently choose set starting paranleter estimates will avoid errors due local minima. The unique contribution paper that it outlines approach for using perspective projection...

Constructing a full CAD model of part requires feature descriptions from all sides; in this case we consider surface patches as the geometric primitives. Most previous research patch extraction has concentrated on extracting single view. This leads to several problems with aligning and combining partial fragments order produce complete models. We have avoided these by adapting our view, range data segmentation program extract patches, thus models, directly fully merged datasets.

The first algorithm to handle a defined class of curved-surface objects is presented. automatically computes the partition Gaussian sphere, and thereby aspect graph, for solids revolution as right, circular, straight, homogeneous generalized cylinders. demonstrates that it possible compute exact graph at least some classes objects. obvious extension this work attempt apply same or similar approach larger more varied objects.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML"...

People often only see what that they already know or explicitly look for. Thus, while mobile users might be interested in the historic background of their surrounding, its current significance relation to specific events, are likely miss places interest if not pointed out. Location based services (LBS) like tourist guides offer a potential technological solution, but production cost multimedia content is prohibitive many cases, limiting coverage such major areas. To provide with information...

Abstract. Matching and fusion of 3D point clouds, such as close range laser scans, is important for creating an integrated model data infrastructure. The Iterative Closest Point algorithm alignment clouds one the most commonly used algorithms matching rigid bodies. Evidently, scans are acquired from different positions might present characterization accuracies, forcing complex data-handling issues. growing demand near real-time applications also introduces new computational requirements...

Abstract. Mobile Mapping is widely used for collecting large amounts of geo-referenced data. An important role plays sensor fusion, in order to evaluate multiple sensors such as laser scanner and cameras jointly. This requires determine the relative orientation between sensors. Based on data a RIEGL VMX-250 mobile mapping system equipped with two scanners, four optional cameras, highly precise GNSS/IMU system, we propose an approach improve camera orientations. A manually determined initial...

Algorithms for computing the aspect graph representation are generalized to include a larger, more realistic domain of objects known as articulated assemblies those composed rigid parts with connections allowed between parts. The generalization suggests two slightly different representations: one that directly summarizes possible general views object and another (hierarchical) form summarizing configurations their respective views. outlined both representations. graphs formed using...

Over the past few years, a number of researchers have presented algorithms for computing aspect graph representation both polyhedra and curved-surface objects. Viewpoint space models included viewing sphere (orthographic projection) all 3-D (perspective projection). The paper presents first (only) complete implementation an algorithm to compute exact class objects using as model viewpoint space. considered is solids revolution defined Right, Circular, Straight, Homogeneous Generalized...

Currently the aspect graph is computed under assumption of perfect resolution in viewpoint, projected image, and object shape. Visual detail represented that an observer might never see practice. By introducing scale into this framework, a mechanism provided for selecting levels are large enough to merit explicit representation, effectively allowing control over size graph. To end space introduced, interpretation dimension terms spatial extent image features considered. A brief example given...

Earth system science is an interdisciplinary effort to understand the fundamentals and interactions of environmental processes. Interdisciplinary research challenging since it demands integration scientific schemes practices from different fields into a collaborative work environment. This paper introduces framework F4ESS that supports this integration. provides methods technologies facilitate development integrative environments for science. enables scientists a) outline structured...

Abstract Planning an event-based monitoring campaign on the regional scale is challenging, e.g. timing and location of visits can dramatically impact efficacy depends optimal environmental conditions required by measurement parameters overarching goal. Therefore we developed a generic planning approach utilizing interactive visualization methods implemented this into component-based web tool called Tocap: Tool for Campaign Planning. As case study, determine most suitable time event-driven,...

A prototype 3D object recognition system is described that composed of three major sections; an representation module, a feature extraction and matching control strategy module. The module uses algorithm which constructs perspective projection aspect graphs convex polyhedra. Fourier descriptors to characterize the complete 2D object. graph application constrained optimization algorithm. implemented in C on Sun workstation, some simple experiments are reported demonstrate validity overall...