基于改进PointRCNN的3D点云目标检测

针对PointRCNN(3D Object Proposal Generation and Detection from Point Cloud)在面对不规则点云时很难提取出有区别特征的问题,提出了一种Point-ANN(3D Object Proposal Generation and Aggregation Neu...     

点云编码综述

点云编码是支撑点云广泛应用的关键技术之一,是近期技术研究和标准化领域的热点。对点云几何信息和属性信息编码技术演进进行了回顾,并针对稠密点云和稀疏点云的几种典型编码方法的编码效率进行了比较。未来点云编码研究将集中于利用帧间预测去除动态点云的不同帧之间的相关性,以及端到端点云编码、任务驱动的点云编码等方面。     

基于点云深度学习的3D目标检测

针对目前利用点云进行3D目标检测的研究较少和检测精度不高的问题,利用Frustum-Pointnets模型实现基于点云的3D目标检测,并在该模型的基础上进行改进,选用不同的激活函数和参数初始化方法进行组合对比,进一步提高模型的精度。实验表明:在选用Swish激活函数和He参数初始化方法时汽车平均检测精度提高了0.31 %,行人平均检测精度提高了0.41 %,骑车人平均检测精度提高了5.5 %。因此改进后的模型能有效提高检测的精度,使得模型能够应用在复杂的场景中。     

基于激光扫描点云的3D地形模拟实验设计

当前3D地形模拟方法未考虑点云数据去噪问题,导致模拟时间较长,准确性较低.为此提出一种基于激光扫描点云的3D地形模拟方法.将噪声点进行分类,去除离群点,通过最小二乘拟合出K邻域内点的最佳逼近平面,判定邻域各点到平面的距离以及设定阈值的大小,以此为依据去除数据中存在的噪声.提取3D地形的轮廓点作为简化点云需要保留的特征点...     

人工智能在点云压缩中的应用前景

点云压缩在沉浸式媒体、自动驾驶等领域有着广泛的应用前景.代表性算法有传统的基于几何、基于映射的编码技术等.随着人工智能在图像压缩、点云处理等课题上取得了巨大的成果,将人工智能应用于点云压缩是一个有潜力的研究方向.本文简要介绍了点云压缩的相关背景,并在此基础上介绍了人工智能在点云压缩中的研究现状,并展望了未来应用前景.     

基于激光三维点云技术的3D景观建模设计

为提升城市园林等类型景观的建模精度,以激光三维点云为技术基础,设计一种三维景观建模方法。采用激光三维点云立体式非接触测量技术,获取景观表面数据点三维坐标,在一个坐标系内统一化各角度点云数据,将顺序点间的最远距离作为滤波标准,设定超过标准点为固定端点,平滑处理图像点云,采用三角形网格参数化策略,映射三维网格模型至二维平面中,取得特征点纹理坐标,利用调和映射算法求解非约束点的纹理坐标,通过自适应部分调整策略,优化点云数据纹理,得到最终的景观模型。试验采集研究区域中一处景观的三维数据,结合景观模型效果与评估指标值得出,所提方法能够有效建立模型,且精准度较高,模型细节信息保存得相对完整。     

怎样将3D图形转为2D卡通

如何在3DS MAX中将三维物件渲染为二维卡通图形,也就是用3D制作出2D,即与专业级相媲美的卡通动画呢?笔者将通过实际的制作经验与大家探讨。 记得在MAX 2.x的时代,就有了很多的插件,例如CartoonRaY、Comicshop和Illustrator等等,都可用来制作卡通效果。最近,德国Cebas公司开发的3DS MAX插件Finalrender又成为CG界MAX迷的炒作热点,但     

3D图形加速器卡向集成多媒体功能方向发展

业界人士正注视着廉价的三维(3D)图形加速器涌入台式多媒体市场,这一趋势始于第一代用于电视游戏的大量生产的3D图形加速器卡的推出,其售价为每块30～50美元,然而,这些卡仅是为中档的电视游戏应用而设计的。随着更先进的芯片组的出现,现能制造开发、能集成许许多多媒体功能的3D图形卡,如DVD译码、MPEG-2重放和AGP能力,以及支持各种高级应用。这些新颖的3D图形卡的报价通常在100～200美元之间,但价格正在下降,由于制造商正在扩大生产能力,因此,今年的价格下降趋势仍将继续。     

三维激光点云下利用Mean_shift的欧式目标分割

随着计算机视觉领域的发展,利用激光点云获取目标三维数据成了当下热点.对目标点云进行有效地分割变得尤为重要.相比于现存的区域生长分割,RANSAC分割与欧式聚类分割等传统的分割算法,提出了Mean-shift的欧式聚类算法.这一算法先对目标点云进行Mean_shift密度聚类粗分割,再结合欧式聚类对粗分割后的点云进行细分...     

改进贪婪投影三角化算法的激光点云快速三维重建

通过三维激光扫描仪获取的点云数据具有密度大、精度高等特点。本文针对贪婪投影三角化算法在对采集的大量点云数据进行三维重建时耗时长,重构的模型表面不够光滑,存在细小孔洞的问题,提出一种改进的点云三维重建算法。该方法首先用体像素网格滤波算法对点云进行下采样;然后使用移动最小二乘算法对输入的点云进行平滑及重采样,并且使用八叉树来代替KD树进行近邻域搜索;最后使用基于移动最小二乘算法的点云法线估计的贪婪投影三角化算法对点云进行重建。经过实验验证,该方法可以缩短重建时间,减少孔洞,并构建出平滑、点云拓扑结构更为准确的模型。     

基于体素化图卷积网络的三维点云目标检测方法

针对激光雷达点云的稀疏性和空间离散分布的特点,通过结合体素划分和图表示方法设计了新的图卷积特征提取模块,提出一种基于体素化图卷积神经网络的激光雷达三维点云目标检测算法。该方法通过消除传统3D卷积神经网络的计算冗余性,不仅提升了网络的目标检测能力,并且提高了点云拓扑信息的分析能力。文中设计的方法在KITTI公开数据集的车辆、行人、骑行者的3D目标检测和鸟瞰图目标检测任务的检测性能相比基准网络均有了有效提升,尤其在车辆3D目标检测任务上最高提升了13.75%。实验表明:该方法采用图卷积特征提取模块有效提高了网络整体检测性能和数据拓扑关系的学习能力,为三维点云目标检测任务提供了新的方法。     

Real-time 3D multi-pedestrian detection and tracking using 3D LiDAR point cloud for mobile robot

Mobile robots are used in modern life; however, object recognition is still insufficient to realize robot navigation in crowded environments. Mobile robots must rapidly and accurately recognize the movements and shapes of pedestrians to navigate safely in pedestrian-rich spaces. This study proposes real-time, accurate, three-dimensional (3D) multi-pedestrian detection and tracking using a 3D light detection and ranging (LiDAR) point cloud in crowded environments. The pedestrian detection quickly segments a sparse 3D point cloud into individual pedestrians using a lightweight convolutional autoencoder and connected-component algorithm. The multi-pedestrian tracking identifies the same pedestrians considering motion and appearance cues in continuing frames. In addition, it estimates pedestrians' dynamic movements with various patterns by adaptively mixing heterogeneous motion models. We evaluate the computational speed and accuracy of each module using the KITTI dataset. We demonstrate that our integrated system, which rapidly and accurately recognizes pedestrian movement and appearance using a sparse 3D LiDAR, is applicable for robot navigation in crowded spaces.     

基于IVCCS的三维点云配准算法

针对传统迭代最近点(ICP)算法在数据丢失以及存在噪声点的情况下配准时间过长、精度较低等问题,提出了一种基于改进的体素云连通性分割(IVCCS)与加权最近邻距离比相结合的配准算法。利用双阈值体素去噪剔除初始种子体素中的噪声体素,解决原本体素云连通性分割算法(VCCS)中因单一约束条件导致种子体素错误剔除的问题,同时将体素云分层去噪来加快配准的运算速度;利用流约束聚类提取点云中的特征点,并依据最近邻距离比验证特征点是否为重合点,赋予不同的权重优化ICP最小目标函数,从而加快配准速度。实验结果表明,该算法相对于传统ICP算法迭代次数减少,在精度与速度方面均有显著提升,相比于基于快速点特征直方图(FPFH)的ICP算法配准精度提高了8.5%~24.7%,速度上提高了65.6%~92.3%,迭代次数减少了16.6%~38%。     

基于注意力机制的三维点云目标识别

针对现有基于深度学习方法的三维点云目标识别算法存在多层感知法缺少点间的特征交互、对点云间欧式距离的依赖、未考虑特征通道层面关联性问题,提出一种基于注意力机制的三维点云(PAttenCls)目标识别算法。采用基于点的空间注意力机制,挖掘各点之间的注意力值,实现自适应的云邻域选择;同时采用基于点的通道注意力机制,给特征通道自适应分配权重,实现特征增强。此外,在网络中添加了一个几何均匀化模块,以应对不同局部区域几何结构的不同特征模式。所提算法在ModelNet40数据集上的识别准确率为93.2%,在ScanObjectNN数据集最难子集上的识别准确率为80.9%,并在实测数据上验证了算法的有效性。实验证明了本文所提算法可以更好地提取点云的特征信息,使点云识别结果更加精准。     

Exploiting color for graph-based 3D point cloud denoising

A point cloud is a representation of a 3D scene as a discrete collection of geometry plus other attributes such as color, normal, transparency associated with each point. The traditional acquisition process of a 3D point cloud, e.g. using depth information acquired directly by active sensors or indirectly from multi-viewpoint images, suffers from a significant amount of noise. Hence, the problem of point cloud denoising has recently received a lot of attention. However, most existing techniques attempt to denoise only the geometry of each point, based on the geometry information of the neighboring points; there are very few works at all considering the problem of denoising the color attributes of a point cloud. In this paper, we move beyond the state of the art and we propose a novel technique employing graph-based optimization, taking advantage of the correlation between geometry and color, and using it as a powerful tool for several different tasks, i.e. color denoising, geometry denoising, and combined geometry and color denoising. The proposed method is based on the notion that the correct location of a point also depends on the color attribute and not only the geometry of the neighboring points, and the correct color also depends on the geometry of the neighbors. The proposed method constructs a suitable k-NN graph from geometry and color and applies graph-based convex optimization to obtain the denoised point cloud. Extensive simulation results on both real-world and synthetic point clouds show that the proposed denoising technique outperforms state-of-the-art methods using both subjective and objective quality metrics.     

3D reconstruction and multiple point cloud registration using a low precision RGB-D sensor

A 3D reconstruction method using feature points is presented and the parameters used to improve the reconstruction are discussed. The precision of the 3D reconstruction is improved by combining point clouds obtained from different viewpoints using structured light. A well-known algorithm for point cloud registration is the ICP (Iterative Closest Point) that determines the rotation and translation that, when applied to one of the point clouds, places both point clouds optimally. The ICP algorithm iteratively executes two main steps: point correspondence determination and registration algorithm. The point correspondence determination is a module that, if not properly executed, can make the ICP converge to a local minimum. To overcome this drawback, two techniques were used. A meaningful set of 3D points using a technique known as SIFT (Scale-invariant feature transform) was obtained and an ICP that uses statistics to generate a dynamic distance and color threshold to the distance allowed between closest points was implemented. The reconstruction precision improvement was implemented using meaningful point clouds and the ICP to increase the number of points in the 3D space. The surface reconstruction is performed using marching cubes and filters to remove the noise and to smooth the surface. The factors that influence the 3D reconstruction precision are here discussed and analyzed. A detailed discussion of the number of frames used by the ICP and the ICP parameters is presented.     

基于激光雷达SLAM的三维点云自适应算法

SLAM(Simultaneously Localization And Mapping)同步定位与地图构建作为移动机器人智能感知的关键技术。但是,大多已有的SLAM方法是在静止环境下实现的,当环境中存在移动频繁的障碍物时,SLAM建图会产生运动畸变,导致机器人无法进行精准的定位导航。同时,激光雷达等三维扫描设备获得的三维点云数据存在着大量的冗余三维数据点,过多的冗余数据不仅浪费大量的存储空间,同时也影响了各种点云处理算法的实时性。针对以上问题,本文提出一种SLAM运动畸变去除方法和一种基于曲率的点云数据分类简化框架。它通过激光插值法优化SLAM运动畸变,将优化后的点云数据分类简化。它能在提高SLAM建图精度,同时也很好的消除三维点云数据中特征不明显区域的冗余数据点,大大提高计算机运行效率。