Raw 3D Point Cloud Data Research Articles

A noise-reduction algorithm for raw 3D point cloud data of asphalt pavement surface texture

High-precision 3D point cloud data have various analyses and application use cases. This study aimed to achieve a more precise noise reduction of the raw 3D point cloud data of asphalt pavements obtained using 3D laser scanning. Hence, a noise-reduction algorithm integrating improved Gaussian filtering and coefficient of variation was developed. A portable laser scanner was used to collect raw, high-precision 3D point cloud data of surface textures from pavement slab samples prepared with three different types of asphalt mixtures: AC-13, SMA-13, and OGFC-13, as well as asphalt from the test sections of the Yakang Expressway. An improved Gaussian filtering and Gaussian filtering that extracts noise using the coefficient of variation were used to filter out the obvious outlier noise and small-scale burr noise, respectively. Finally, the filtering effect of the proposed algorithm, Gaussian filtering, median filtering, and mean filtering on raw 3D point cloud data of pavement textures was evaluated through subjective visual quality and objective index evaluations. The results showed that the proposed algorithm filters out noise while preserving the micro-texture structure information, outperforming Gaussian filtering, median filtering, and mean filtering.

Comparison of 3D Point Clouds Obtained by Terrestrial Laser Scanning and Personal Laser Scanning on Forest Inventory Sample Plots

In forest inventory, trees are usually measured using handheld instruments; among the most relevant are calipers, inclinometers, ultrasonic devices, and laser range finders. Traditional forest inventory has been redesigned since modern laser scanner technology became available. Laser scanners generate massive data in the form of 3D point clouds. We have developed a novel methodology to provide estimates of the tree positions, stem diameters, and tree heights from these 3D point clouds. This dataset was made publicly accessible to test new software routines for the automatic measurement of forest trees using laser scanner data. Benchmark studies with performance tests of different algorithms are welcome. The dataset contains co-registered raw 3D point-cloud data collected on 20 forest inventory sample plots in Austria. The data were collected by two different laser scanning systems: (1) A mobile personal laser scanner (PLS) (ZEB Horizon, GeoSLAM Ltd., Nottingham, UK) and (2) a static terrestrial laser scanner (TLS) (Focus3D X330, Faro Technologies Inc., Lake Mary, FL, USA). The data also contain digital terrain models (DTMs), field measurements as reference data (ground-truth), and the output of recent software routines for the automatic tree detection and the automatic stem diameter measurement.

3D-SiamRPN: An End-to-End Learning Method for Real-Time 3D Single Object Tracking Using Raw Point Cloud

3D single object tracking is a key issue for autonomous following robot, where the robot should robustly track and accurately localize the target for efficient following. In this paper, we propose a 3D tracking method called 3D-SiamRPN Network to track a single target object by using raw 3D point cloud data. The proposed network consists of two subnetworks. The first subnetwork is feature embedding subnetwork which is used for point cloud feature extraction and fusion. In this subnetwork, we first use PointNet++ to extract features of point cloud from template and search branches. Then, to fuse the information of features in the two branches and obtain their similarity, we propose two cross correlation modules, named Pointcloud-wise and Point-wise respectively. The second subnetwork is region proposal network(RPN), which is used to get the final 3D bounding box of the target object based on the fusion feature from cross correlation modules. In this subnetwork, we utilize the regression and classification branches of a region proposal subnetwork to obtain proposals and scores, thus get the final 3D bounding box of the target object. Experimental results on KITTI dataset show that our method has a competitive performance in both Success and Precision compared to the state-of-the-art methods, and could run in real-time at 20.8 FPS. Additionally, experimental results on H3D dataset demonstrate that our method also has good generalization ability and could achieve good tracking performance in a new scene without re-training.

A Point-Based and Image-Based Multi-Pass Rendering Technique for Visualizing Massive 3D Point Clouds in VR Environments

Real-time rendering for 3D point clouds allows for interactively exploring and inspecting real-world assets, sites, or regions on a broad range of devices but has to cope with their vastly different computing capabilities. Virtual reality (VR) applications rely on high frame rates (i.e., around 90 fps as opposed to 30 - 60 fps) and show high sensitivity to any kind of visual artifacts, which are typical for 3D point cloud depictions (e.g., holey surfaces or visual clutter due to inappropriate point sizes). We present a novel rendering system that allows for an immersive, nausea-free exploration of arbitrary large 3D point clouds on state-of-the-art VR devices such as HTC Vive and Oculus Rift. Our approach applies several point-based and image-based rendering techniques that are combined using a multipass rendering pipeline. The approach does not require to derive generalized, mesh-based representations in a preprocessing step and preserves precision and density of the raw 3D point cloud data. The presented techniques have been implemented and evaluated with massive real-world data sets from aerial, mobile, and terrestrial acquisition campaigns containing up to 2.6 billion points to show the practicability and scalability of our approach.

Survey on 3D Surface Reconstruction

The recent advent of increasingly affordable and powerful 3D scanning devices capable of capturing high resolution range data about real-world objects and environments has fueled research into effective 3D surface reconstruction techniques for rendering the raw point cloud data produced by many of these devices into a form that would make it usable in a variety of application domains. This paper, therefore, provides an overview of the existing literature on surface reconstruction from 3D point clouds. It explains some of the basic surface reconstruction concepts, describes the various factors used to evaluate surface reconstruction methods, highlights some commonly encountered issues in dealing with the raw 3D point cloud data and delineates the tradeoffs between data resolution/accuracy and processing speed. It also categorizes the various techniques for this task and briefly analyzes their empirical evaluation results demarcating their advantages and disadvantages. The paper concludes with a cross-comparison of methods which have been evaluated on the same benchmark data sets along with a discussion of the overall trends reported in the literature. The objective is to provide an overview of the state of the art on surface reconstruction from point cloud data in order to facilitate and inspire further research in this area.

Sensing and responding to the changes of geometric surfaces in flexible manufacturing and assembly

A flexible manufacturing system is capable of dealing with changes and uncertainties occurring in a manufacturing environment. A rigid automation system adapts changes through reconfiguration or manual modification of its control programmes. In contrast, an advanced system responds to the changes autonomously or with a minimal manual intervention. One critical technology for a flexible manufacturing system is to detect or predict the changes and uncertainties and take them into account for the control of the manufacturing system. In this paper, an intelligent manufacturing system is proposed to accommodate product geometric variants autonomously by integrating a 3D vision system into manufacturing processes. The proposed system is capable of acquiring vision sensor data, detecting the geometric changes, and modifying the control programmes of the machine in response to the changes. In this innovative work, a new method for converting the raw vision sensor data into a surface model of the product is presented. A new surface reconstruction algorithm is proposed to generate the surface model directly from the raw 3D point-cloud data of a product. In addition, an advanced programming algorithm that uses acquired surface models to automatically update robot programmes for a new task with minimal manual intervention is presented.