Point Cloud Processing Algorithms Research Articles

A Novel 3-D Path Extraction Method for Arc Welding Robot Based on Stereo Structured Light Sensor

With the rapid development of computer vision and industrial technology, the requirements of the intelligent welding robots are increasing in the real industrial production. The traditional teaching-playback mode and the off-line programming mode cannot meet the automation demand and self-adaptive ability of welding robots. In order to improve the efficiency of welding robots, this paper proposes a novel 3-D path extraction method of weld seams based on a stereo-structured light sensor. Faced with the low efficiency of the line-structured light and the poor robustness of passive vision, the seam extraction based on a point cloud processing algorithm is proposed which could well adapt to the weld seams with different types and different groove sizes. Meanwhile, the position information and pose information of weld seam are established to serve 3-D path teaching of a welding robot. The experimental results show that the maximum path extraction error of V-type butt joint is less than 0.7 mm. The proposed scheme could well serve for the 3-D path teaching task before welding.

Practical shape analysis and segmentation methods for point cloud models

Current point cloud processing algorithms do not have the capability to automatically extract semantic information from the observed scenes, except in very specialized cases. Furthermore, existing mesh analysis paradigms cannot be directly employed to automatically perform typical shape analysis tasks directly on point cloud models.We present a potent framework for shape analysis, similarity and segmentation of noisy point cloud models for real objects of engineering interest, models that may be incomplete. The proposed framework relies on spectral methods and the heat diffusion kernel to construct compact shape signatures, and we show that the framework supports a variety of clustering techniques that have traditionally been applied only on mesh models. We developed and implemented one practical and convergent estimate of the Laplace–Beltrami operator for point clouds as well as a number of clustering techniques adapted to work directly on point clouds to produce geometric features of engineering interest. The key advantage of this framework is that it supports practical shape analysis capabilities that operate directly on point cloud models of objects without requiring surface reconstruction or global meshing. We show that the proposed technique is robust against typical noise present in possibly incomplete point clouds, and segment point clouds scanned by depth cameras (e.g. Kinect) into semantically-meaningful sub-shapes.

Unsupervised detection of vineyards by 3D point-cloud UAV photogrammetry for precision agriculture

An effective management of precision viticulture processes relies on robust crop monitoring procedures and, in the near future, to autonomous machine for automatic site-specific crop managing. In this context, the exact detection of vineyards from 3D point-cloud maps, generated from unmanned aerial vehicles (UAV) multispectral imagery, will play a crucial role, e.g. both for achieve enhanced remotely sensed data and to manage path and operation of unmanned vehicles.In this paper, an innovative unsupervised algorithm for vineyard detection and vine-rows features evaluation, based on 3D point-cloud maps processing, is presented. The main results are the automatic detection of the vineyards and the local evaluation of vine rows orientation and of inter-rows spacing.The overall point-cloud processing algorithm can be divided into three mains steps: (1) precise local terrain surface and height evaluation of each point of the cloud, (2) point-cloud scouting and scoring procedure on the basis of a new vineyard likelihood measure, and, finally, (3) detection of vineyard areas and local features evaluation.The algorithm was found to be efficient and robust: reliable results were obtained even in the presence of dense inter-row grassing, many missing plants and steep terrain slopes. Performances of the algorithm were evaluated on vineyard maps at different phenological phase and growth stages. The effectiveness of the developed algorithm does not rely on the presence of rectilinear vine rows, being also able to detect vineyards with curvilinear vine row layouts.

Sector Based Linear Regression, a New Robust Method for the Multiple Linear Regression

This paper describes a new robust multiple linear regression method, which based on the segmentation of the N dimensional space to N+1 sector. An N dimensional regression plane is located so that the half (or other) part of the points are under this plane in each sector. This article also presents a simple algorithm to calculate the parameters of this regression plane. This algorithm is scalable well by the dimension and the count of the points, and capable to calculation with other (not 0.5) quantiles. This paper also contains some studies about the described method, which analyze the result with different datasets and compares to the linear least squares regression. Sector Based Linear Regression (SBLR) is the multidimensional generalization of the mathematical background of a point cloud processing algorithm called Fitting Disc method, which has been already used in practice to process LiDAR data. A robust regression method can be used also in many other fields.

Robust normal estimation and region growing segmentation of infrastructure 3D point cloud models

Modern remote sensing technologies such as three-dimensional (3D) laser scanners and image-based 3D scene reconstruction are in increasing demand for applications in civil infrastructure design, maintenance, operation, and as-built construction verification. The complex nature of the 3D point clouds these technologies generate, as well as the often massive scale of the 3D data, make it inefficient and time consuming to manually analyze and manipulate point clouds, and highlights the need for automated analysis techniques. This paper presents one such technique, a new region growing algorithm for the automated segmentation of both planar and non-planar surfaces in point clouds. A core component of the algorithm is a new point normal estimation method, an essential task for many point cloud processing algorithms. The newly developed estimation method utilizes robust multivariate statistical outlier analysis for reliable normal estimation in complex 3D models, considering that these models often contain regions of varying surface roughness, a mixture of high curvature and low curvature regions, and sharp features. An adaptation of Mahalanobis distance, in which the mean vector and covariance matrix are derived from a high-breakdown multivariate location and scale estimator called Deterministic MM-estimator (DetMM) is used to find and discard outlier points prior to estimating the best local tangent plane around any point in a cloud. This approach is capable of more accurately estimating point normals located in highly curved regions or near sharp features. Thereafter, the estimated point normals serve a region growing segmentation algorithm that only requires a single input parameter, an improvement over existing methods which typically require two control parameters. The reliability and robustness of the normal estimation subroutine was compared against well-known normal estimation methods including the Minimum Volume Ellipsoid (MVE) and Minimum Covariance Determinant (MCD) estimators, along with Maximum Likelihood Sample Consensus (MLESAC). The overall region growing segmentation algorithm was then experimentally validated on several challenging 3D point clouds of real-world infrastructure systems. The results indicate that the developed approach performs more accurately and robustly in comparison with conventional region growing methods, particularly in the presence of sharp features, outliers and noise.

3D Reconstruction of Specific Target Based on Kinectfusion

A new 3D reconstruction method of specific target which is mixed in complex scene is proposed by making full use of the point cloud model acquired by KinectFusion and other point cloud processing algorithm. The target needed to be reconstructed can be obtained by using point cloud segmentation based on RANSAC and Euclidean cluster extraction as well as using surface reconstruction based on concave hull. The purpose of segmentation is to break the point cloud down into distinct clusters which can then be processed independently. The surface reconstruction is to generate a surface of the target that is extracted by segmentation. Experimental results show that the method in this paper has an outstanding performance in 3D reconstruction of the specific target.

HELIOS: A MULTI-PURPOSE LIDAR SIMULATION FRAMEWORK FOR RESEARCH, PLANNING AND TRAINING OF LASER SCANNING OPERATIONS WITH AIRBORNE, GROUND-BASED MOBILE AND STATIONARY PLATFORMS

In many technical domains of modern society, there is a growing demand for fast, precise and automatic acquisition of digital 3D models of a wide variety of physical objects and environments. Laser scanning is a popular and widely used technology to cover this demand, but it is also expensive and complex to use to its full potential. However, there might exist scenarios where the operation of a <i>real</i> laser scanner could be replaced by a computer simulation, in order to save time and costs. This includes scenarios like teaching and training of laser scanning, development of new scanner hardware and scanning methods, or generation of artificial scan data sets to support the development of point cloud processing and analysis algorithms. To test the feasibility of this idea, we have developed a highly flexible laser scanning simulation framework named <i>Heidelberg LiDAR Operations Simulator (HELIOS)</i>. HELIOS is implemented as a Java library and split up into a core component and multiple extension modules. Extensible Markup Language (XML) is used to define scanner, platform and scene models and to configure the behaviour of modules. Modules were developed and implemented for (1) loading of simulation assets and configuration (i.e. 3D scene models, scanner definitions, survey descriptions etc.), (2) playback of XML survey descriptions, (3) TLS survey planning (i.e. automatic computation of recommended scanning positions) and (4) interactive real-time 3D visualization of simulated surveys. As a proof of concept, we show the results of two experiments: First, a survey planning test in a scene that was specifically created to evaluate the quality of the survey planning algorithm. Second, a simulated TLS scan of a crop field in a precision farming scenario. The results show that HELIOS fulfills its design goals.

HELIOS: A MULTI-PURPOSE LIDAR SIMULATION FRAMEWORK FOR RESEARCH, PLANNING AND TRAINING OF LASER SCANNING OPERATIONS WITH AIRBORNE, GROUND-BASED MOBILE AND STATIONARY PLATFORMS

Abstract. In many technical domains of modern society, there is a growing demand for fast, precise and automatic acquisition of digital 3D models of a wide variety of physical objects and environments. Laser scanning is a popular and widely used technology to cover this demand, but it is also expensive and complex to use to its full potential. However, there might exist scenarios where the operation of a real laser scanner could be replaced by a computer simulation, in order to save time and costs. This includes scenarios like teaching and training of laser scanning, development of new scanner hardware and scanning methods, or generation of artificial scan data sets to support the development of point cloud processing and analysis algorithms. To test the feasibility of this idea, we have developed a highly flexible laser scanning simulation framework named Heidelberg LiDAR Operations Simulator (HELIOS). HELIOS is implemented as a Java library and split up into a core component and multiple extension modules. Extensible Markup Language (XML) is used to define scanner, platform and scene models and to configure the behaviour of modules. Modules were developed and implemented for (1) loading of simulation assets and configuration (i.e. 3D scene models, scanner definitions, survey descriptions etc.), (2) playback of XML survey descriptions, (3) TLS survey planning (i.e. automatic computation of recommended scanning positions) and (4) interactive real-time 3D visualization of simulated surveys. As a proof of concept, we show the results of two experiments: First, a survey planning test in a scene that was specifically created to evaluate the quality of the survey planning algorithm. Second, a simulated TLS scan of a crop field in a precision farming scenario. The results show that HELIOS fulfills its design goals.