Methodology for Extraction of Tunnel Cross-Sections Using Dense Point Cloud Data

Tunnel deformation monitoring is a crucial task to evaluate tunnel stability during the metro operation period.Terrestrial Laser Scanning(TLS)can collect high density and high accuracy point cloud data in a few minutes as an innovation technique,which provides promising applications in tunnel deformation monitoring.Here,an efficient method for extracting tunnel cross-sections and convergence analysis using dense TLS point cloud data is proposed.First,the tunnel orientation is determined using principal component analysis(PCA)in the Euclidean plane.Two control points are introduced to detect and remove the unsuitable points by using point cloud division and then the ground points are removed by defining an elevation value width of 0.5 m.Next,a z-score method is introduced to detect and remove the outlies.Because the tunnel cross-section’s standard shape is round,the circle fitting is implemented using the least-squares method.Afterward,the convergence analysis is made at the angles of 0°,30°and 150°.The proposed approach’s feasibility is tested on a TLS point cloud of a Nanjing subway tunnel acquired using a FARO X330 laser scanner.The results indicate that the proposed methodology achieves an overall accuracy of 1.34 mm,which is also in agreement with the measurements acquired by a total station instrument.The proposed methodology provides new insights and references for the applications of TLS in tunnel deformation monitoring,which can also be extended to other engineering applications.

Indoor Space Modeling and Parametric Component Construction Based on 3D Laser Point Cloud Data

In order to enhance modeling efficiency and accuracy,we utilized 3D laser point cloud data for indoor space modeling.Point cloud data was obtained with a 3D laser scanner and optimized with Autodesk Recap and Revit software to extract geometric information about the indoor environment.Furthermore,we proposed a method for constructing indoor elements based on parametric components.The research outcomes of this paper will offer new methods and tools for indoor space modeling and design.The approach of indoor space modeling based on 3D laser point cloud data and parametric component construction can enhance modeling efficiency and accuracy,providing architects,interior designers,and decorators with a better working platform and design reference.

Automated Rock Detection and Shape Analysis from Mars Rover Imagery and 3D Point Cloud Data

A new object-oriented method has been developed for the extraction of Mars rocks from Mars rover data. It is based on a combination of Mars rover imagery and 3D point cloud data. First, Navcam or Pancam images taken by the Mars rovers are segmented into homogeneous objects with a mean-shift algorithm. Then, the objects in the segmented images are classified into small rock candidates, rock shadows, and large objects. Rock shadows and large objects are considered as the regions within which large rocks may exist. In these regions, large rock candidates are extracted through ground-plane fitting with the 3D point cloud data. Small and large rock candidates are combined and postprocessed to obtain the final rock extraction results. The shape properties of the rocks （angularity, circularity, width, height, and width-height ratio） have been calculated for subsequent ~eological studies.

Development of vehicle-recognition method on water surfaces using LiDAR data:SPD^(2)(spherically stratified point projection with diameter and distance)

Swarm robot systems are an important application of autonomous unmanned surface vehicles on water surfaces.For monitoring natural environments and conducting security activities within a certain range using a surface vehicle,the swarm robot system is more efficient than the operation of a single object as the former can reduce cost and save time.It is necessary to detect adjacent surface obstacles robustly to operate a cluster of unmanned surface vehicles.For this purpose,a LiDAR(light detection and ranging)sensor is used as it can simultaneously obtain 3D information for all directions,relatively robustly and accurately,irrespective of the surrounding environmental conditions.Although the GPS(global-positioning-system)error range exists,obtaining measurements of the surface-vessel position can still ensure stability during platoon maneuvering.In this study,a three-layer convolutional neural network is applied to classify types of surface vehicles.The aim of this approach is to redefine the sparse 3D point cloud data as 2D image data with a connotative meaning and subsequently utilize this transformed data for object classification purposes.Hence,we have proposed a descriptor that converts the 3D point cloud data into 2D image data.To use this descriptor effectively,it is necessary to perform a clustering operation that separates the point clouds for each object.We developed voxel-based clustering for the point cloud clustering.Furthermore,using the descriptor,3D point cloud data can be converted into a 2D feature image,and the converted 2D image is provided as an input value to the network.We intend to verify the validity of the proposed 3D point cloud feature descriptor by using experimental data in the simulator.Furthermore,we explore the feasibility of real-time object classification within this framework.

Classification of rice seed variety using point cloud data combined with deep learning

Rice variety selection and quality inspection are key links in rice planting.Compared with two-dimensional images,three-dimensional information on rice seeds shows the appearance characteristics of rice seeds more comprehensively and accurately.This study proposed a rice variety classification method using three-dimensional point cloud data of the surface of rice seeds combined with a deep learning network to achieve the rapid and accurate identification of rice varieties.First,a point cloud collection platform was set up with a Raytrix light field camera as the core to collect three-dimensional point cloud data on the surface of rice seeds;then,the collected point cloud was filled,filtered and smoothed;after that,the point cloud segmentation is based on the RANSAC algorithm,and the point cloud downsampling is based on a combination of random sampling algorithm and voxel grid filtering algorithm.Finally,the processed point cloud was input to the improved PointNet network for feature extraction and species classification.The improved PointNet network added a cross-level feature connection structure,made full use of features at different levels,and better extracted the surface structure features of rice seeds.After testing,the improved PointNet model had an average classification accuracy of 89.4%for eight varieties of rice,which was 1.2%higher than that of the PointNet model.The method proposed in this study combined deep learning and point cloud data to achieve the efficient classification of rice varieties.

Classified denoising method for laser point cloud data of stored grain bulk surface based on discrete wavelet threshold

Surfaces of stored grain bulk are often reconstructed from organized point sets with noise by 3-D laser scanner in an online measuring system.As a result,denoising is an essential procedure in processing point cloud data for more accurate surface reconstruction and grain volume calculation.A classified denoising method was presented in this research for noise removal from point cloud data of the grain bulk surface.Based on the distribution characteristics of cloud point data,the noisy points were divided into three types:The first and second types of the noisy points were either sparse points or small point cloud data deviating and suspending from the main point cloud data,which could be deleted directly by a grid method;the third type of the noisy points was mixed with the main body of point cloud data,which were most difficult to distinguish.The point cloud data with those noisy points were projected into a horizontal plane.An image denoising method,discrete wavelet threshold(DWT)method,was applied to delete the third type of the noisy points.Three kinds of denoising methods including average filtering method,median filtering method and DWT method were applied respectively and compared for denoising the point cloud data.Experimental results show that the proposed method remains the most of the details and obtains the lowest average value of RMSE(Root Mean Square Error,0.219)as well as the lowest relative error of grain volume(0.086%)compared with the other two methods.Furthermore,the proposed denoising method could not only achieve the aim of removing noisy points,but also improve self-adaptive ability according to the characteristics of point cloud data of grain bulk surface.The results from this research also indicate that the proposed method is effective for denoising noisy points and provides more accurate data for calculating grain volume.

Accuracy of common stem volume formulae using terrestrial photogrammetric point clouds:a case study with savanna trees in Benin

Recent applications of digital photogrammetry in forestry have highlighted its utility as a viable mensuration technique.However,in tropical regions little research has been done on the accuracy of this approach for stem volume calculation.In this study,the performance of Structure from Motion photogrammetry for estimating individual tree stem volume in relation to traditional approaches was evaluated.We selected 30 trees from five savanna species growing at the periphery of the W National Park in northern Benin and measured their circumferences at different heights using traditional tape and clinometer.Stem volumes of sample trees were estimated from the measured circumferences using nine volumetric formulae for solids of revolution,including cylinder,cone,paraboloid,neiloid and their respective fustrums.Each tree was photographed and stem volume determined using a taper function derived from tri-dimensional stem models.This reference volume was compared with the results of formulaic estimations.Tree stem profiles were further decomposed into different portions,approximately corresponding to the stump,butt logs and logs,and the suitability of each solid of revolution was assessed for simulating the resulting shapes.Stem volumes calculated using the fustrums of paraboloid and neiloid formulae were the closest to reference volumes with a bias and root mean square error of 8.0%and 24.4%,respectively.Stems closely resembled fustrums of a paraboloid and a neiloid.Individual stem portions assumed different solids as follows:fustrums of paraboloid and neiloid were more prevalent from the stump to breast height,while a paraboloid closely matched stem shapes beyond this point.Therefore,a more accurate stem volumetric estimate was attained when stems were considered as a composite of at least three geometric solids.

基于改进PointNet++的输电线路关键部位点云语义分割研究

输电线路的关键部位包括塔身、导线、绝缘子、避雷线以及引流线,无人机精细化导航的首要任务是构造输电线路的点云地图并从中分割出上述部位。为解决现有算法在输电线路的绝缘子、引流线等精细结构分割时精度低的问题,通过改进PointNet++算法,提出了一种面向输电线路精细结构的点云分割方法。首先,基于无人机机载激光雷达在现场采集的点云数据,构造了输电线路点云分割数据集;其次,通过对比实验,筛选出在本输电线路场景下合理的数据增强方法,并对数据集进行了数据增强;最后,将自注意力机制以及倒置残差结构和PointNet++相结合,设计了输电线路关键部位点云语义分割算法。实验结果表明:该改进PointNet++算法在全场景输电线路现场点云数据作为输入的前提下,首次实现了对引流线、绝缘子等输电线路中精细结构和导线、杆塔塔身以及输电线路无关背景点的同时分割,平均交并比(mean intersection over union,mIoU)达80.79%,所有类别分割的平均F_(1)值(F1 score)达88.99%。

ALGORITHM OF PRETREATMENT ON AUTOMOBILE BODY POINT CLOUD

As point cloud of one whole vehicle body has the traits of large geometric dimension, huge data and rigorous reverse precision, one pretreatment algorithm on automobile body point cloud is put forward. The basic idea of the registration algorithm based on the skeleton points is to construct the skeleton points of the whole vehicle model and the mark points of the separate point cloud, to search the mapped relationship between skeleton points and mark points using congruence triangle method and to match the whole vehicle point cloud using the improved iterative closed point （ICP） algorithm. The data reduction algorithm, based on average square root of distance, condenses data by three steps, computing datasets＇ average square root of distance in sampling cube grid, sorting order according to the value computed from the first step, choosing sampling percentage. The accuracy of the two algorithms above is proved by a registration and reduction example of whole vehicle point cloud of a certain light truck.

三维激光扫描点云数据在CloudWorx for MicroStation下的处理

三维激光扫描技术作为一种先进的测量手段应用前景十分广阔,但是,在应用其扫描所得的点云数据进行内处理上又遇到了许多技术性的问题。CAD系统处理大量点云数据过程中存在局限性,一旦用CAD系统处理点云时,CAD程序就会出现错误操作提示,甚至完全停止进程。CloudWorx克服了这些局限性,它避免了直接输入数据,而是利用Cyclone技术作为MicroStation环境下有效管理和解决点云的工具,使MicroStation内点云操作与CAD程序执行再无冲突。本文主要介绍了并且还介绍了CloudWorx模块的功能,并举例就点云数据在CloudWorx for MicroStation软件环境下的处理工作进行了详细的介绍。

Multi-view ladar data registration in obscure environment

Multi-view laser radar （ladar） data registration in obscure environments is an important research field of obscured target detection from air to ground. There are few overlap regions of the observational data in different views because of the occluder, so the multi-view data registration is rather difficult. Through indepth analyses of the typical methods and problems, it is obtained that the sequence registration is more appropriate, but needs to improve the registration accuracy. On this basis, a multi-view data registration algorithm based on aggregating the adjacent frames, which are already registered, is proposed. It increases the overlap region between the pending registration frames by aggregation and further improves the registration accuracy. The experiment results show that the proposed algorithm can effectively register the multi-view ladar data in the obscure environment, and it also has a greater robustness and a higher registration accuracy compared with the sequence registration under the condition of equivalent operating efficiency.

PointNet的点云数据集的破损测试与深度解读

目前人们对二维图像的研究已经取得了非常好的结果,然而随着深度学习的发展,研究正在逐步由二维向三维数据发展,并且应用领域越来越广泛,比如自动驾驶领域的三维场景建模,VR虚拟环境等。对三维数据的研究也逐渐实现了由有序输入到无序输入的过度并且取得了很高的成绩。Point Net则是第一个突破点云数据无序性输入的深度神经网络,值得人们深入的研究和借鉴。但是目前对破损和遮挡的点云数据问题还有待研究。本文着重对Point Net进行了深入研究并对点云数据进行了攻击和测试。测试发现当点云数据dropout约75%以后物体识别准确率显著下降,overlap12. 5%以后准确率也下降了近4个点,值得后续深入的研究和攻克。

基于PointNet的点云数据处理及识别技术应用研究

主要基于PointNet模型对点云数据进行特征分类处理,直接对无序化的三维点云数据进行无规则输入处理,通过对无序化的点云数据中的每一个点进行单独的处理,来实现点云数据的输入,与点的输入顺序没有关系,在PointNet中最重要的方法是对称函数最大池化来合并点云数据中每一个点的信息,输出分类特征集合或分割结果,然后结合计算机视觉的神经网络和深度学习等方法,来理解和分析点云数据对物体和环境的智能化识别的应用研究。

基于Point-to-Plane ICP的点云与影像数据自动配准

针对三维激光点云与二维影像数据的融合问题,采用了一种基于Point-to-Plane ICP的配准方法;该方法仅采用一块普通的平面黑白棋盘格作为标定板,能同时完成单目相机的标定与三维激光扫描仪和相机的联合标定,进而实现三维点云数据与二维影像数据的配准;与以往基于控制点或者边缘对应的配准方法不同,该方法使用RANSAC算法自动提取场景中的标定平面,通过优化点到平面的距离来求取两组数据的变换。实验结果表明,该配准方法减少了人工的干预,并获得了很高的精度。

改进的3D-BoNet算法应用于点云实例分割与三维重建

为了更好地利用点云数据重建室内三维模型,本文提出了一种基于3D-BoNet-IAM算法的室内场景三维重建方法。该方法通过改进3D-BoNet算法提高点云数据的实例分割精度。针对点云数据缺失问题,提出了基于平面基元合并优化的拟合平面方法,利用拟合得到的新平面重建建筑表面模型。在S3DIS和ScanNet V2数据集上验证3D-BoNet算法的改进效果。试验结果表明,本文提出的3D-BoNet-IAM算法比原始算法分割精度提高了3.3%;对比本文建模效果与其他建模效果发现,本文方法的建模效果更准确。本文方法能够提高室内点云数据的实例分割精度,同时得到高质量的室内三维模型。

K-means聚类精简点云驱动PointNet++的行星齿轮故障诊断

复杂装备的三维模型点云数据具有非结构化、无序性、离散性的特点,数据精简策略和深度神经网络模型构建被视为点云数据驱动的机械设备故障诊断关键技术难点。提出了一种K-means聚类(K均值聚类算法)精简点云驱动PointNet++的行星齿轮故障诊断方法。首先,提出了基于K-means的点云数据精简策略实现了在充分保留细节特征的前提下,精简84%的冗余数据;其次,构建了简度、速度、精度的精简效果三维评价指标体系并对精简算法进行评价;最后,构建了能够提取局部特征的PointNet++故障诊断模型。实验结果表明,相比于点云数据直接驱动PointNet++,K-means聚类精简点云驱动PointNet++的行星齿轮故障诊断的准确率提升了6.9%,表明了所提方法的有效性。

基于深度学习的移动机器人语义SLAM方法研究

为了给移动机器人提供细节丰富的三维语义地图,支撑机器人的精准定位,本文提出一种结合RGB-D信息与深度学习结果的机器人语义同步定位与建图方法。改进了ORB-SLAM2算法的框架,提出一种可以构建稠密点云地图的视觉同步定位与建图系统;将深度学习的目标检测算法YOLO v5与视觉同步定位与建图系统融合,反映射为三维点云语义标签,结合点云分割完成数据关联和物体模型更新,并用八叉树的地图形式存储地图信息;基于移动机器人平台,在实验室环境下开展移动机器人三维语义同步定位与建图实验,实验结果验证了本文语义同步定位与建图算法的语义信息映射、点云分割与语义信息匹配以及三维语义地图构建的有效性。

基于车载点云的道路三维实景建模方法研究

传统的基础测绘存在组织管理固化、服务模式落后、产品形式单一等问题,在新型基础测绘体系下形成了全要素三维实景模型这一成果。本文探讨基于车载点云进行城市道路三维实景建模方法研究,并以某城市主干路为试验对象,对道路及道路两侧部件点云数据进行矢量化得到道路全要素地形数据,以部件点云数据为参考结合外业调绘尺寸用3ds Max软件制作道路部件模板库,并结合点云数据和矢量数据对各类要素进行单体化,最后将道路模型和部件模型融合。结果表明,基于车载点云数据构建的城市道路全要素实景模型不仅可以保证场景的完整性和真实性,还减少了作业时间和成本,实现了各类模型之间的无缝结合,制作完成的模型精度也能满足项目精度要求。

地面激光扫描点云与无人机影像点云融合应用

通过建立高精度的桥梁三维点云模型,检查桥梁病害情况并拟合绘制出桥梁线形。首先以无人机近景摄影、环绕飞行、井字飞行获取某双线特大桥梁主体与细部纹理数据,然后将不同航线采集的数据在Context Capture软件里面进行三维重建,将桥梁主体与细部影像融合生成完整桥梁点云1。运用Trimble SX12仪器完成对桥梁一体化扫描,获得完整桥梁点云2。提出基于双向KD-tree优化的ICP(Iterative Closest Point)算法对无人机航摄桥梁点云1与地面激光扫描桥梁点云数据2进行配准融合,加密后的桥梁点云用于建立运营铁路双线特大桥精细化三维实景建模。提出基于KD-tree的PCA(Principal Component Analysis)算法完整提取出桥梁吊索点云,运用最小二乘法拟合出桥梁拱轴线线形、RANSAC算法拟合出桥面线形。通过与单一无人机、单一地面激光扫描精度及完整性对比分析,以验证融合建模的有效性。研究结果表明:融合建模的模型水平精度1.71 cm、垂直方向精度1.25 cm,较单一无人机建模精度在水平与竖直方向分别提升16.59%与20.89%;融合建模的完整性为98.17%,纹理效果更加真实,并检查出桥墩存在蜂窝麻面、渗水等病害,拱肋存在涂装锈蚀、破裂等病害。该研究可为桥梁三维点云模型应用研究提供思路参考,具有较好的应用前景。

三维激光扫描技术在历史建筑测绘中的应用——以闽清县历史建筑测绘建档项目为例

历史建筑作为城市的文脉,承载着一座城市的历史,受各种因素影响遭受不断的侵蚀甚至灭失,其保护形势越来越严峻。文章以闽清县历史建筑保护测绘为例,采用三维激光扫描技术,结合无人机倾斜摄影技术,对历史建筑真彩色三维点云模型建设进行了探索研究,为历史建筑的测绘资料建档和文物保护工作积累了宝贵的技术经验。结语对该技术进行了总结分析,认为三维激光扫描技术不仅能大幅提升工作效率和测量成果精度,还可实现历史建筑三维可视化,具有广阔的应用前景。