- 标题
- 摘要
- 关键词
- 实验方案
- 产品
-
A Geometry-Based Point Cloud Reduction Method for Mobile Augmented Reality System
摘要： In this paper, a geometry-based point cloud reduction method is proposed, and a real-time mobile augmented reality system is explored for applications in urban environments. We formulate a new objective function which combines the point reconstruction errors and constraints on spatial point distribution. Based on this formulation, a mixed integer programming scheme is utilized to solve the points reduction problem. The mobile augmented reality system explored in this paper is composed of the offline and online stages. At the offline stage, we build up the localization database using structure from motion and compress the point cloud by the proposed point cloud reduction method. While at the online stage, we compute the camera pose in real time by combining an image-based localization algorithm and a continuous pose tracking algorithm. Experimental results on benchmark and real data show that compared with the existing methods, this geometry-based point cloud reduction method selects a point cloud subset which helps the image-based localization method to achieve higher success rate. Also, the experiments conducted on a mobile platform show that the reduced point cloud not only reduces the time consuming for initialization and re-initialization, but also makes the memory footprint small, resulting a scalable and real-time mobile augmented reality system.
关键词： augmented reality,point cloud reduction,structure from motion,mobile platform
更新于2025-09-10 09:29:36
-
Development of deep learning architecture for automatic classification of outdoor mobile LiDAR data
摘要： This paper proposes a deep convolutional neural network (CNN) architecture for automatic classification of mobile laser scanning (MLS) data obtained for outdoor environment, which are characterized by noise, clutter, large size and larger quantum of information. The developed architecture introduces a look up table (LUT) based approach, which retains the geometry of the input MLS point cloud while rescaling. Further, with the voxelisation of the input MLS sample, the ambiguity of selecting one out of multiple point values within a voxel is resolved. The performance of the architecture is evaluated on MLS data of outdoor environment in two instances, first using tree and non-tree classes (non-tree class has objects like electric pole, wire, low vegetation, wall, house and ground) and then with tree and electric pole classes. Additional testing is carried out by mixing the outdoor MLS data of tree and electric pole classes with three classes of indoor objects, taken from Modelnet dataset, thereby assessing the architecture efficacy over an ensemble of three-dimensional (3D) datasets. Classification of tree and non-tree classes, followed by tree and electric pole classes from MLS samples result in total accuracies of 86.0%, 90.0% respectively and kappa values of 72.0%, 78.7% respectively. Moreover, for the combinations of MLS and Modelnet classes, the classification results are promising, reaching a total accuracy of 95.2% and kappa of 92.5%. The LUT based approach has shown better classification over the traditional rescaling approach for the MLS dataset, resulting in an enhancement up to 9.0% and 18.0% in total accuracy and kappa, respectively. With different varieties of tree, non-tree and electric pole samples, the proposed architecture has shown its potential for automatic classification of MLS data with high accuracy. This study further reveals that the accuracy of classification is improved by introducing more spatial features in the input layer. The accuracies produced in this work can be further improved with the availability of better hardware resources.
关键词： outdoor environment,deep learning,mobile laser scanning,point cloud,convolutional neural network,classification
更新于2025-09-10 09:29:36
-
[IEEE 2018 International Conference on Communication and Signal Processing (ICCSP) - Chennai, India (2018.4.3-2018.4.5)] 2018 International Conference on Communication and Signal Processing (ICCSP) - Estimation of Peaks and Canopy Height Using LiDAR Data
摘要： 3D city modelling plays a major role in today’s smart city development. LiDAR facilitates acquisition and processing of point cloud data. The proposed plan in this paper will estimate the height of a building. LiDAR data processing is done by the software, LAS tools. In this paper the study focuses on LiDAR scan of urban scenes (buildings). Canopy Height Model (CHM) of a building is obtained which is useful in finding the height of a building, shape of the building and peak location. This can be a contribution to smart city modeling and urban development. Canopy height model for LiDAR data is obtained through classifiers and filters available in LAStools. Also this paper presents about Digital Terrain Model (DTM) and Digital Surface Model (DSM) for LiDAR point cloud data. Further the obtained CHM can be used for the application of finding the number of buildings in an area which are above a threshold height.
关键词： LAStools,DSM,Point cloud data,DTM,LiDAR scan,Canopy height model
更新于2025-09-09 09:28:46
-
From LiDAR Waveforms to Hyper Point Clouds: A Novel Data Product to Characterize Vegetation Structure
摘要： Full waveform (FW) LiDAR holds great potential for retrieving vegetation structure parameters at a high level of detail, but this prospect is constrained by practical factors such as the lack of available handy processing tools and the technical intricacy of waveform processing. This study introduces a new product named the Hyper Point Cloud (HPC), derived from FW LiDAR data, and explores its potential applications, such as tree crown delineation using the HPC-based intensity and percentile height (PH) surfaces, which shows promise as a solution to the constraints of using FW LiDAR data. The results of the HPC present a new direction for handling FW LiDAR data and offer prospects for studying the mid-story and understory of vegetation with high point density (~182 points/m2). The intensity-derived digital surface model (DSM) generated from the HPC shows that the ground region has higher maximum intensity (MAXI) and mean intensity (MI) than the vegetation region, while having lower total intensity (TI) and number of intensities (NI) at a given grid cell. Our analysis of intensity distribution contours at the individual tree level exhibit similar patterns, indicating that the MAXI and MI decrease from the tree crown center to the tree boundary, while a rising trend is observed for TI and NI. These intensity variable contours provide a theoretical justification for using HPC-based intensity surfaces to segment tree crowns and exploit their potential for extracting tree attributes. The HPC-based intensity surfaces and the HPC-based PH Canopy Height Models (CHM) demonstrate promising tree segmentation results comparable to the LiDAR-derived CHM for estimating tree attributes such as tree locations, crown widths and tree heights. We envision that products such as the HPC and the HPC-based intensity and height surfaces introduced in this study can open new perspectives for the use of FW LiDAR data and alleviate the technical barrier of exploring FW LiDAR data for detailed vegetation structure characterization.
关键词： vegetation structure,HPC-based intensity surface,percentile height,tree segmentation,gridding,hyper point cloud (HPC),full waveform LiDAR
更新于2025-09-09 09:28:46
-
3D Building Roof Modeling by Optimizing Primitive’s Parameters Using Constraints from LiDAR Data and Aerial Imagery
摘要： In this paper, a primitive-based 3D building roof modeling method, by integrating LiDAR data and aerial imagery, is proposed. The novelty of the proposed modeling method is to represent building roofs by geometric primitives and to construct a cost function by using constraints from both LiDAR data and aerial imagery simultaneously, so that the accuracy potential of the different sensors can be tightly integrated for the building model generation by an integrated primitive’s parameter optimization procedure. To verify the proposed modeling method, both simulated data and real data with simple buildings provided by ISPRS (International Society for Photogrammetry and Remote Sensing), were used in this study. The experimental results were evaluated by the ISPRS, which demonstrate the proposed modeling method can integrate LiDAR data and aerial imagery to generate 3D building models with high accuracy in both the horizontal and vertical directions. The experimental results also show that by adding a component, such as a dormer, to the primitive, a variant of the simple primitive is constructed, and the proposed method can generate a building model with some details.
关键词： building modeling,optimization,primitive,LiDAR,aerial imagery,model-based,data fusion,point cloud
更新于2025-09-09 09:28:46
-
High Precision Individual Tree Diameter and Perimeter Estimation from Close-Range Photogrammetry
摘要： Close-range photogrammetry (CRP) can be used to provide precise and detailed three-dimensional data of objects. For several years, CRP has been a subject of research in forestry. Several studies have focused on tree reconstruction at the forest stand, plot, and tree levels. In our study, we focused on the reconstruction of trees separately within the forest stand. We investigated the influence of camera lens, tree species, and height of diameter on the accuracy of the tree perimeter and diameter estimation. Furthermore, we investigated the variance of the perimeter and diameter reference measurements. We chose four tree species (Fagus sylvatica L., Quercus petraea (Matt.) Liebl., Picea abies (L.) H. Karst. and Abies alba Mill.). The perimeters and diameters were measured at three height levels (0.8 m, 1.3 m, and 1.8 m) and two types of lenses were used. The data acquisition followed a circle around the tree at a 3 m radius. The highest accuracy of the perimeter estimation was achieved when a fisheye lens was used at a height of 1.3 m for Fagus sylvatica (root mean square error of 0.25 cm). Alternatively, the worst accuracy was achieved when a non-fisheye lens was used at 1.3 m for Quercus petraea (root mean square error of 1.27 cm). The tree species affected the estimation accuracy for both diameters and perimeters.
关键词： close-range photogrammetry,fisheye lens,trunk perimeter,circle fitting,trunk diameter,convex hull,point cloud
更新于2025-09-09 09:28:46
-
Improving LiDAR classification accuracy by contextual label smoothing in post-processing
摘要： We propose a contextual label-smoothing method to improve the LiDAR classification accuracy in a post-processing step. Under the framework of global graph-structured regularization, we enhance the effectiveness of label smoothing from two aspects. First, each point can collect sufficient label-relevant neighborhood information to verify its label based on an optimal graph. Second, the input label probability set is improved by probabilistic label relaxation to be more consistent with the spatial context. With this optimal graph and reliable label probability set, the final labels are computed by graph-structured regularization. We demonstrate the contextual label-smoothing approach on two separate urban airborne LiDAR datasets with complex urban scenes. Significant improvements in the classification accuracies are achieved without losing small objects (such as fa?ades and cars). The overall accuracy is increased by 7.01% on the Vienna dataset and 6.88% on the Vaihingen dataset. Moreover, most large, wrongly labeled regions are corrected by long-range interactions that are derived from the optimal graph, and misclassified regions that lack neighborhood communications in terms of correct labels are also corrected with the probabilistic label relaxation.
关键词： Neighborhood dependency,Optimal neighborhood,Point cloud,Probabilistic label relaxation
更新于2025-09-04 15:30:14
-
[IEEE 2018 Global Smart Industry Conference (GloSIC) - Chelyabinsk, Russia (2018.11.13-2018.11.15)] 2018 Global Smart Industry Conference (GloSIC) - Three-Dimensional Workpiece Scanning for Technological Operation Parameters Determination in Turbine Rotor Restoration
摘要： A progressive method of turbine rotor journal restoration is machining with a direct location on a bearing bottom half. In the process, the work surface form depends on the type of rotor journal previous wear. In the previous research it was assumed that the rotor journal had the form of a cylindrical surface with a polyhedral generating line. It has been established technological operation parameters determination, for example, the tool axis installation angle, significantly depends on a form of the specified generating line. However, the real surface of a rotor journal can differ from the specified noncircular cylinder. Thus, a problem of technological operation appropriate parameters determination for various rotor journal surfaces is vital. One of the known ways of obtaining exact object surface data is the formation of a point cloud by means of a 3D scanner. The scanning result is the digital double of a workpiece surface in the form of the point cloud. So, the development of the machined surface circularity deviation calculation method, based on coordinates of a point cloud and technological operation parameters, is a relevant problem. Such methodology has been developed and forecasting of the most rational parameters of technological operation to minimize the surface circularity deviation has been conducted.
关键词： digital double of surface,machining of locating surface,point cloud,3D scanning,Turbine rotor journal restoration
更新于2025-09-04 15:30:14
-
Comprehensive Remote Sensing || Geometric Processing: Active Sensor Modeling and Calibration (LiDAR)
摘要： The main concept of laser ranging is facilitating the estimation of the distance between the laser beam firing point and its footprint. For time-of-flight systems (ToF), whether pulse-based or phase-shift-based systems, the range is evaluated through the estimation of the time lapse between laser emission and reception. This time when coupled with the speed of the laser beam provides an estimate of the distance between the laser beam firing point and its footprint. One should note that this chapter is not concerned with triangulation-based ranging systems that are comprised of an integrated laser and camera system—that is, the range is estimated using an accurately estimated relative relationship between the laser and camera units.
关键词： system calibration,LiDAR,point cloud,laser ranging,time-of-flight
更新于2025-09-04 15:30:14
-
[Institution of Engineering and Technology 12th European Conference on Antennas and Propagation (EuCAP 2018) - London, UK (9-13 April 2018)] 12th European Conference on Antennas and Propagation (EuCAP 2018) - Increased reliability of outdoor millimeter-wave link simulations by leveraging lidar point cloud
摘要： Exploitation of the millimeter-wave (mmWave) band for both the small-cell radio access (RAN) and in-street backhauling must contribute to strongly increase the capacity of next-generation outdoor wireless networks. Beside the numerous technological challenges faced by the equipment manufacturers, the definition of practical deployment requirements and elaboration of efficient network design procedures are key issues for future network providers and the refinement of business models. The work reported in the present paper is part of research activity on small-cell and wireless backhaul design methodologies. The authors analyze how digital geographical data and radio propagation models may be enhanced and tailored to answer new mmWave simulation requirements. Three different propagation methods are compared in an urban 60 GHz scenario: based on free building database, traditional high-resolution map data, or Lidar point cloud. The advantage of the multi-paths prediction versus direct-path is also investigated, using the Lidar point cloud as a simulation input.
关键词： millimeter-wave propagation,Lidar point cloud,simulation
更新于2025-09-04 15:30:14