The degree of spatial similarity plays an important role in map generalization, yet there has been no quantitative research into it. To fill this gap, this study first defines map scale change and spatial similarity d...The degree of spatial similarity plays an important role in map generalization, yet there has been no quantitative research into it. To fill this gap, this study first defines map scale change and spatial similarity degree/relation in multi-scale map spaces and then proposes a model for calculating the degree of spatial similarity between a point cloud at one scale and its gener- alized counterpart at another scale. After validation, the new model features 16 points with map scale change as the x coordinate and the degree of spatial similarity as the y coordinate. Finally, using an application for curve fitting, the model achieves an empirical formula that can calculate the degree of spatial similarity using map scale change as the sole independent variable, and vice versa. This formula can be used to automate algorithms for point feature generalization and to determine when to terminate them during the generalization.展开更多
The purpose of reverse engineering is to convert a large point cloud into a CAD model. In reverse engineering, the key issue is segmentation, i.e. studying how to subdivide the point cloud into smaller regions, where ...The purpose of reverse engineering is to convert a large point cloud into a CAD model. In reverse engineering, the key issue is segmentation, i.e. studying how to subdivide the point cloud into smaller regions, where each of them can be approximated by a single surface. Segmentation is relatively simple, if regions are bounded by sharp edges and small blends; problems arise when smoothly connected regions need to be separated. In this paper, a modified self-organizing feature map neural network （SOFM） is used to solve segmentation problem. Eight dimensional feature vectors （3-dimensional coordinates, 3-dimensional normal vectors, Gaussian curvature and mean curvature） are taken as input for SOFM. The weighted Euclidean distance measure is used to improve segmentation result. The method not only can deal with regions bounded by sharp edges, but also is very efficient to separating smoothly connected regions. The segmentation method using SOFM is robust to noise, and it operates directly on the point cloud. An examples is given to show the effect of SOFM algorithm.展开更多
针对动态物体严重干扰同时定位与建图(SLAM)系统正常运行的问题,提出一种基于目标检测和特征点关联的动态视觉SLAM算法。首先,利用YOLOv5目标检测网络得到环境中潜在动态物体的信息,并基于简易目标跟踪对图像漏检进行补偿;其次,为解决...针对动态物体严重干扰同时定位与建图(SLAM)系统正常运行的问题,提出一种基于目标检测和特征点关联的动态视觉SLAM算法。首先,利用YOLOv5目标检测网络得到环境中潜在动态物体的信息,并基于简易目标跟踪对图像漏检进行补偿;其次,为解决单一特征点的几何约束方法易出现误判的问题,依据图像的位置信息和光流信息建立特征点关联,再结合极线约束判断关系网的动态性;再次,结合两种方法剔除图像中的动态特征点,并用剩余的静态特征点加权估计位姿;最后,对静态环境建立稠密点云地图。在TUM(Technical University of Munich)公开数据集上的对比和消融实验的结果表明,与ORB-SLAM2和DS-SLAM(Dynamic Semantic SLAM)相比,所提算法在高动态场景下的绝对轨迹误差(ATE)中的均方根误差(RMSE)分别至少降低了95.22%和5.61%。可见,所提算法在保证实时性的同时提高了准确性和鲁棒性。展开更多
同时定位与建图(Simultaneous localization and mapping,SLAM)能够在未知环境中构建地图并为机器人提供定位信息,是移动机器人领域重要研究方向之一.当前,大多数SLAM算法在静态环境中有较好的表现,但是在车辆和行人等运动物体较多的环...同时定位与建图(Simultaneous localization and mapping,SLAM)能够在未知环境中构建地图并为机器人提供定位信息,是移动机器人领域重要研究方向之一.当前,大多数SLAM算法在静态环境中有较好的表现,但是在车辆和行人等运动物体较多的环境中,广泛存在的动态点使激光点云前后帧的配准精度不高,降低了动态场景下定位和建图的准确性.针对激光点云中存在动态点的问题,本文对SLAM的前端特征提取及后端回环检测模块分别进行改进,以去除动态点,提升SLAM在动态环境下的性能.针对SLAM前端,提出了一种分步的地面分割方法,依据点云高度信息完成地面点粗提取以矫正点云,再使用随机采样一致性方法对矫正后的点云进行精细的地面分割,最后根据高度阈值采用种子生长聚类方法提取非地面动态点,并进行特征提取与配准;针对SLAM后端,使用点云描述子替代传统方法中基于空间位置关系的回环检测方法,以减小累计误差、提高回环检测灵敏度.实验结果显示,本方法在M2DGR street_08序列数据集上较现有方法均方根误差最大降低29.8%,在KITTI04序列数据集上均方根误差最大降幅达42.7%,说明本方法能有效增强动态环境下SLAM系统的全局一致性与定位精度.展开更多
近年来,随着我国煤矿业的快速发展,智能化技术的运用越来越广泛。其中,露天煤矿环境的精确定位导航技术研发显得尤为重要。同步定位和地图构建(Simultaneous Localization and Mapping,SLAM)作为无人驾驶的关键技术,在露天煤矿中的应用...近年来,随着我国煤矿业的快速发展,智能化技术的运用越来越广泛。其中,露天煤矿环境的精确定位导航技术研发显得尤为重要。同步定位和地图构建(Simultaneous Localization and Mapping,SLAM)作为无人驾驶的关键技术,在露天煤矿中的应用面临诸多挑战。由于露天煤矿道路周围环境特征点较少,且环境退化严重,SLAM技术需要根据稀疏的特征点进行定位和地图构建,难度较大。此外,由于斜坡和道路不平,传感器易产生抖动,导致机器人运行时的运动畸变问题。针对这些问题,文中提出了一种新的解决方案。首先,对传感器外部参数进行重新标定,采用惯导和激光雷达融合的方式,以增强数据的一致性和准确性。在此基础上,采用全特征点匹配方式,直接对激光雷达采集的数据进行点云降采样提取。通过在算法前端对预处理后的激光点云数据添加迭代最近点(Iterative Closest Point,ICP)匹配提取出关键帧点云X,再结合惯导数据对点云信息进行畸变校正形成点云P,再次通过迭代最近点配准X和P。此外,后端采用因子图加入了回环检测提高约束的方法,进一步提高算法在露天煤矿环境下的定位精度和建图效果。试验结果表明,文中所提算法具有较高的定位精度和完整的建图效果,未产生明显的畸变。侧壁纹理清晰,具有一定的鲁棒性,有效提高了在露天煤矿环境下的鲁棒性和精度。展开更多
基金funded by the Natural Science Foundation Committee,China(41364001,41371435)
文摘The degree of spatial similarity plays an important role in map generalization, yet there has been no quantitative research into it. To fill this gap, this study first defines map scale change and spatial similarity degree/relation in multi-scale map spaces and then proposes a model for calculating the degree of spatial similarity between a point cloud at one scale and its gener- alized counterpart at another scale. After validation, the new model features 16 points with map scale change as the x coordinate and the degree of spatial similarity as the y coordinate. Finally, using an application for curve fitting, the model achieves an empirical formula that can calculate the degree of spatial similarity using map scale change as the sole independent variable, and vice versa. This formula can be used to automate algorithms for point feature generalization and to determine when to terminate them during the generalization.
基金Supported by the National Natural Science Foundation of China(60573177), the Aeronautical Science Foundation of China (04H53059) , the natural Science Foundation of Henan Province (200510078010) and Youth Science Foundation at North China Institute of Water Conservancy and Hydroelectric Power(HSQJ2004003)
文摘The purpose of reverse engineering is to convert a large point cloud into a CAD model. In reverse engineering, the key issue is segmentation, i.e. studying how to subdivide the point cloud into smaller regions, where each of them can be approximated by a single surface. Segmentation is relatively simple, if regions are bounded by sharp edges and small blends; problems arise when smoothly connected regions need to be separated. In this paper, a modified self-organizing feature map neural network （SOFM） is used to solve segmentation problem. Eight dimensional feature vectors （3-dimensional coordinates, 3-dimensional normal vectors, Gaussian curvature and mean curvature） are taken as input for SOFM. The weighted Euclidean distance measure is used to improve segmentation result. The method not only can deal with regions bounded by sharp edges, but also is very efficient to separating smoothly connected regions. The segmentation method using SOFM is robust to noise, and it operates directly on the point cloud. An examples is given to show the effect of SOFM algorithm.
文摘针对动态物体严重干扰同时定位与建图(SLAM)系统正常运行的问题,提出一种基于目标检测和特征点关联的动态视觉SLAM算法。首先,利用YOLOv5目标检测网络得到环境中潜在动态物体的信息,并基于简易目标跟踪对图像漏检进行补偿;其次,为解决单一特征点的几何约束方法易出现误判的问题,依据图像的位置信息和光流信息建立特征点关联,再结合极线约束判断关系网的动态性;再次,结合两种方法剔除图像中的动态特征点,并用剩余的静态特征点加权估计位姿;最后,对静态环境建立稠密点云地图。在TUM(Technical University of Munich)公开数据集上的对比和消融实验的结果表明,与ORB-SLAM2和DS-SLAM(Dynamic Semantic SLAM)相比,所提算法在高动态场景下的绝对轨迹误差(ATE)中的均方根误差(RMSE)分别至少降低了95.22%和5.61%。可见,所提算法在保证实时性的同时提高了准确性和鲁棒性。
文摘同时定位与建图(Simultaneous localization and mapping,SLAM)能够在未知环境中构建地图并为机器人提供定位信息,是移动机器人领域重要研究方向之一.当前,大多数SLAM算法在静态环境中有较好的表现,但是在车辆和行人等运动物体较多的环境中,广泛存在的动态点使激光点云前后帧的配准精度不高,降低了动态场景下定位和建图的准确性.针对激光点云中存在动态点的问题,本文对SLAM的前端特征提取及后端回环检测模块分别进行改进,以去除动态点,提升SLAM在动态环境下的性能.针对SLAM前端,提出了一种分步的地面分割方法,依据点云高度信息完成地面点粗提取以矫正点云,再使用随机采样一致性方法对矫正后的点云进行精细的地面分割,最后根据高度阈值采用种子生长聚类方法提取非地面动态点,并进行特征提取与配准;针对SLAM后端,使用点云描述子替代传统方法中基于空间位置关系的回环检测方法,以减小累计误差、提高回环检测灵敏度.实验结果显示,本方法在M2DGR street_08序列数据集上较现有方法均方根误差最大降低29.8%,在KITTI04序列数据集上均方根误差最大降幅达42.7%,说明本方法能有效增强动态环境下SLAM系统的全局一致性与定位精度.
