多镜头组合式相机的全景SLAM

同步定位与地图构建(simultaneous localization and mapping,SLAM)是摄影测量、计算机视觉和机器人学的研究热点,并广泛应用于移动测图系统、机器人、无人驾驶车等。本文提出一种针对多镜头组合式全景相机的基于特征的SLAM解决方案。首先,本文建立了鱼眼相机的高精度检校模型,以保证鱼眼相机与全景相机之间的高精度坐标转换;然后,将多镜头组合式全景成像模型嵌入SLAM的初始化、局部地图生成、关键帧选取、图优化等各个流程中。此外,考虑全景相机变形大、基线长的不利因素,本文在特征匹配、平差、特征点跟踪等SLAM的各个步骤都进行了针对性改进。本文在两套车载全景数据集共8000余张全景影像上进行试验。结果表明,本文所提出的全景SLAM很好地实现了全景相机的自动定位与地图构建功能,并达到了接近GPS参考的极高定位精度而无须借助GPS/IMU组合导航系统。相对于主流的基于平面相机的各类SLAM系统,如Mono-SLAM、Stereo-SLAM以及RGB-DSLAM,本文提出的全景SLAM可作为良好的补充,并为GPS信号失锁时的传感器定位提供廉价的自动解决方案。     

Validation of geometric models for fisheye lenses

The paper focuses on the photogrammetric investigation of geometric models for different types of optical fisheye constructions (equidistant, equisolid-angle, sterographic and orthographic projection). These models were implemented and thoroughly tested in a spatial resection and a self-calibrating bundle adjustment. For this purpose, fisheye images were taken with a Nikkor 8 mm fisheye lens on a Kodak DSC 14n Pro digital camera in a hemispherical calibration room. Both, the spatial resection and the bundle adjustment resulted in a standard deviation of unit weight of 1/10 pixel with a suitable set of simultaneous calibration parameters introduced into the camera model. The camera-lens combination was treated with all of the four basic models mentioned above. Using the same set of additional lens distortion parameters, the differences between the models can largely be compensated, delivering almost the same precision parameters. The relative object space precision obtained from the bundle adjustment was ca. 1:10 000 of the object dimensions. This value can be considered as a very satisfying result, as fisheye images generally have a lower geometric resolution as a consequence of their large field of view and also have a inferior imaging quality in comparison to most central perspective lenses.     

利用鱼眼相机对人群进行运动估计

人群运动估计是人群行为分析的重要步骤。特定场景的人群运动分析和监控,是维护公共安全和社会稳定的一个必要措施,也是视频监控领域的一个研究难点。利用鱼眼相机视场大、无视觉盲区的优点,提出了一种基于特征点光流的人群运动估计方法。首先,采用一种基于面积反馈机制的混合高斯背景差分方法,对原始视频图像进行预处理,并利用圆拟合的方法获取兴趣区域;其次,为了在保证准确描述人群目标的同时提高算法的实时性,提出一种基于边缘密度非均匀采样的人群特征点提取方法来描述运动的人群目标,并利用Lucas & Kanade光流法计算光流场;最后,为了解决远近人群的尺寸大小不一致的问题和鱼眼相机的畸变问题,采用鱼眼相机的透视加权模型,计算人群运动加权统计直方图,获取人群在鱼眼图像中的全局运动方向和速度。实验结果表明,针对密集的人群,该方法能有效、实时地估计人群的运动方向和速度,为人群行为分析提供有力的研究基础。     

消费级无人机在大比例尺测图中应用与精度评价

以分别搭载大视场角普通相机和鱼眼相机的消费级无人机为例,在对相机的高精度标定基础上,选取典型丘陵和山地实验区分别对两种消费级无人机大比例尺测图能力进行实验和精度评价,结果证明消费级无人机在合理的航测方式下,其测图结果的空三精度、分辨率及高程精度均可满足国家1∶500大比例尺地形测图的精度要求,可以推广应用到小范围大比例尺测图和修测任务中。     

彩色点云匀色的Wallis算法

通过可见光相机与三维激光扫描仪之间的精确标定,可以生成具有颜色信息的点云,在文物建模、移动测图、地下空间调查等领域具有广泛应用。点云的色彩质量依赖可见光相机的成像质量,受硬件水平、光照条件、拍摄参数等诸多因素干扰,经常出现偏暗、过曝、不均匀等现象,然而目前缺乏点云色彩匀色方法的研究。因此,本文将Wallis匀色算法从二维拓展到三维,利用模板影像的颜色信息,改善三维点云的色彩效果。实验证明,三维Wallis匀色算法可以有效改善点云的波段比例、亮度与对比度。     

多像位姿估计的全景纹理映射算法

纹理映射技术作为获取具有丰富纹理信息的真彩色点云的有效手段,正以其独特的优势广泛地应用于众多行业领域。研究了一种利用三维激光扫描仪与外置数码相机联合标定解算多张影像位姿并获取全景真彩色点云的方法。其基本思想是利用摄像机与激光扫描仪固有的相对位置姿态,通过对首张影像进行标定得到其位置姿态后,利用摄像机空间旋转的几何特性,根据首张影像的位姿获取其余影像的位姿,继而完成多张影像的纹理映射,获取全景彩色点云。对比目前主流的全景影像纹理映射算法,该算法在精度与效率上均有一定提高。对多种点云数据进行纹理映射实验,结果表明,该方法能够快速准确地获取真三维全景彩色点云,为三维精细化建模提供了数据基础。     

最小割与深度学习联合优化的室内粘连点云分割方法

随着数字城市的发展,城市三维模型重建对三维点云结构化的需求与精度要求越来越高。如何有效准确地分割室内语义模型与三维重构是当前研究的热点问题。点云分割分类是室内点云结构化的重要基础,如何将粘连点云构件进行准确分割并用于室内点云结构化,是当前城市建模的难点。本文提出了一种面向室内粘连点云数据的分割分类方法。首先,利用深度学习网络处理室内点云数据;其次,对点云数据进行标签分类,得到目标标签点云;然后,利用欧氏算法对目标点云进行聚类分割,通过室内语义构件包围盒信息计算各目标中心点坐标与水平半径;最后,利用点云最小割实现室内粘连点云的准确分割。利用3组室内场景中获取的数据对分割方法的精度及有效性进行了验证。结果表明,该分割优化方法具有较高的精度与数据完整性。     

Crowd sensing and spatiotemporal analysis in urban open space using multi-viewpoint geotagged videos

Increasing concern for urban public safety has motivated the deployment of a large number of surveillance cameras in open spaces such as city squares, stations, and shopping malls. The efficient detection of crowd dynamics in urban open spaces using multi-viewpoint surveillance videos continues to be a fundamental problem in the field of urban security. The use of existing methods for extracting features from video images has resulted in significant progress in single-camera image space. However, surveillance videos are geotagged videos with location information, and few studies have fully exploited the spatial semantics of these videos. In this study, multi-viewpoint videos in geographic space are used to fuse object trajectories for crowd sensing and spatiotemporal analysis. The YOLOv3-DeepSORT model is used to detect a pedestrian and extract the corresponding image coordinates, combine spatial semantics (such as the positions of the pedestrian in the field of view of the camera) to build a projection transformation matrix and map the object recorded by a single camera to geographic space. Trajectories from multi-viewpoint videos are fused based on the features of location, time, and directions to generate a complete pedestrian trajectory. Then, crowd spatial pattern analysis, density estimation, and motion trend analysis are performed. Experimental results demonstrate that the proposed method can be used to identify crowd dynamics and analyze the corresponding spatiotemporal pattern in an urban open space from a global perspective, providing a means of intelligent spatiotemporal analysis of geotagged videos.     

RGB-D SLAM在室内高精度三维测图中的应用

高精度三维测图是室内三维制图的重要支撑,基于三维激光雷达扫描技术的三维测图成本高,需要提前布置标靶,在室内复杂环境中易导致数据不完整;基于图像序列的三维重建建模时间长,易受多种因素影响。针对以上问题,本文将RGB-D SLAM技术应用于室内高精度三维测图中。通过将深度相机与SLAM技术相结合,计算相机位姿并恢复三维空间信息,获取室内三维点云模型,并以目标物实际量测为基准评价密集点云精度。试验结果表明,该方法可快速获取精度较高的三维点云模型,成本低且效率高,能够较好地满足应用需求。     

基于深度残差网络的机载LiDAR点云分类

机载LiDAR点云的分类是利用其进行城市场景三维重建的关键步骤之一。为充分利用现有的图像领域性能较好的深度学习网络模型,提高点云分类精度,并降低训练时间和对训练样本数量的要求,本文提出一种基于深度残差网络的机载LiDAR点云分类方法。首先提取归一化高程、表面变化率、强度和归一化植被指数4种具有较高区分度的点云低层次特征;然后通过设置不同的邻域大小和视角,利用所提出的点云特征图生成策略,得到多尺度和多视角点云特征图;再将点云特征图输入到预训练的深度残差网络,提取多尺度和多视角深层次特征;最后构建并训练神经网络分类器,利用训练的模型对待分类点云进行预测,经后处理得到分类结果。利用ISPRS三维语义标记竞赛的公开标准数据集进行试验,结果表明,本文方法可有效区分建筑物、地面、车辆等8类地物,分类结果的总体精度为87.1%,可为城市场景三维重建提供可靠的信息。     

利用车载LiDAR点云数据提取城市道路边界

随着高精地图产业的兴起,精确提取道路边界点云数据成为研究的重点.本文首先将车载LiDAR扫描系统获取的城市道路数据根据采集轨迹进行分段,对每一段路段点云进行滤波处理;然后通过分析点云的高程与平面信息,采用点云分割算法分离路面与非路面点云,再对处理后的路面点云进行投影;最后运用边界特征估计提取算法获取道路边界点云.通过对...     

车载移动测量系统装备研制与应用

为实现空间信息的快速及高效获取,在集成GNSS、IMU、全景相机、激光雷达等主要传感器的基础上,研制了车载全景激光耦合式移动测量系统。根据精密解算的位置姿态信息,对全景影像和激光点云进行了耦合,生成含有真彩色点云的可量测实景影像,提出鸟瞰实景影像、基于浏览器的海量点云网络发布等点云应用新形式。该系统在渝武高速公路扩能改造工程快速移动测量中的成功应用,证明系统不仅安装简便、采集快速、处理高效、成果丰富,而且具备厘米级的测量精度,有效提高了空间信息的获取及处理效率。     

硐室摄影地质编录原理及其精度

阐述了摄影地质编录原理及其系统工作流程，并研究了硐室的摄影地质编录的精度，研究表明，硐室数字影像展示图足以满足作为编录底图的精度要求，摄影测量能够达到测定地质状要素的精度要求。     

A geospatial hybrid cloud platform based on multi-sourced computing and model resources for geosciences

Cloud computing has been considered as the next-generation computing platform with the potential to address the data and computing challenges in geosciences. However, only a limited number of geoscientists have been adapting this platform for their scientific research mainly due to two barriers: 1) selecting an appropriate cloud platform for a specific application could be challenging, as various cloud services are available and 2) existing general cloud platforms are not designed to support geoscience applications, algorithms and models. To tackle such barriers, this research aims to design a hybrid cloud computing (HCC) platform that can utilize and integrate the computing resources across different organizations to build a unified geospatial cloud computing platform. This platform can manage different types of underlying cloud infrastructure (e.g., private or public clouds), and enables geoscientists to test and leverage the cloud capabilities through a web interface. Additionally, the platform also provides different geospatial cloud services, such as workflow as a service, on the top of common cloud services (e.g., infrastructure as a service) provided by general cloud platforms. Therefore, geoscientists can easily create a model workflow by recruiting the needed models for a geospatial application or task on the fly. A HCC prototype is developed and dust storm simulation is used to demonstrate the capability and feasibility of such platform in facilitating geosciences by leveraging across-organization computing and model resources.     

利用图割算法进行城市密集点云表面模型重建

利用倾斜影像获得的密集点云来构建表面模型是基于倾斜影像进行三维重建的核心之一。本文针对现行密集点云表面模型重建存在的建模效率低、表面选取不真实等问题,提出了一种基于图割算法的城市密集点云表面模型重建方法。利用该方法重建城市密集点云表面模型,首先通过预处理软件对无人机倾斜影像进行空中三角测量,并利用空中三角测量的解算结果生成密集点云;然后对密集点云添加相应的边,同时对三维点云根据距离进行选取合并;最后根据三维点云形成的四面体和三角面建立图割问题,并通过求解图割问题来求取最优的密集点云表面模型。为证明这种方法的可行性和有效性,使用城市地区的无人机倾斜影像数据进行城市密集点云表面模型重建,试验结果表明,该方法具有可行性好、建模效果好、处理速度快等优势。     

A bottom-up approach to segment individual deciduous trees using leaf-off lidar point cloud data

Light Detection and Ranging (Lidar) can generate three-dimensional (3D) point cloud which can be used to characterize horizontal and vertical forest structure, so it has become a popular tool for forest research. Recently, various methods based on top-down scheme have been developed to segment individual tree from lidar data. Some of these methods, such as the one developed by Li et al. (2012), can obtain the accuracy up to 90% when applied in coniferous forests. However, the accuracy will decrease when they are applied in deciduous forest because the interlacing tree branches can increase the difficulty to determine the tree top. In order to solve challenges of the tree segmentation in deciduous forests, we develop a new bottom-up method based on the intensity and 3D structure of leaf-off lidar point cloud data in this study. We applied our algorithm to segment trees in a forest at the Shavers Creek Watershed in Pennsylvania. Three indices were used to assess the accuracy of our method: recall, precision and F-score. The results show that the algorithm can detect 84% of the tree (recall), 97% of the segmented trees are correct (precision) and the overall F-score is 90%. The result implies that our method has good potential for segmenting individual trees in deciduous broadleaf forest.     

无人机大比例尺航测系统的研制及应用

以无人机作为低空遥感平台,将选择的轻小型量测相机与定制的小型双频高动态GPS进行搭载集成,并自主研制了一种三轴稳定云台装置,可保持无人机航摄时相机姿态的稳定,使获得的航空影像满足大比例尺测图的规范要求。将该系统应用于1：500大比例尺地形图测绘项目,结果表明该系统的成图精度完全可以满足国家航测内业成图规范及城市测量规范要求。     

利用波形信息的加权曲面拟合LiDAR点云滤波

为发挥机载全波形激光探测与测量（light detection and ranging,LiDAR）技术优势,提高数字高程模型（digital elevation model,DEM）生成精度,提出了一种利用波形信息的加权曲面拟合LiDAR点云滤波方法。该方法利用全局收敛LM解算离散点云与波形参数,引入波形信息与抗差估计原理检测异常种子点,依据波形参数对地形曲面进行加权拟合,综合考虑滤波窗口尺寸与曲面拟合中误差影响设置自适应高差阈值。选取中国黑河综合遥感联合实验中的城市区域、耕地区域与山地区域数据进行实验,结果表明,相比传统方法,所提方法的波形分解结果更加可靠,点云滤波精度进一步提高,具备较高实用价值。     

北京市土地资源和建设用地综合决策分析平台

本文通过对建设用地资源潜力评价体系的研究,建立了适用于北京市的建设用地资源潜力评价体系:制定了"先筛选,后评价"的建设用地评价思路和工作流程。建立了可利用建设用地筛选模型;在理论构建的基础上,综合利用GIS技术,开发建立了北京市土地资源和建设用地综合决策分析平台。研究发现只有保障了空间数据质量,可利用建设用地资源潜力评价的准确性和可持续性才能得到根本保证。     

面向水库大坝形变监测的三维激光扫描点云分析方法

根据黄龙带水库大坝两期高精度三维激光大坝扫描点云数据,本文提出基于正态分布的点云数据配准算法、改进原始渐进加密三角网滤波算法,运用点到面形变量对比分析模式监测大坝形变量,并进行源程序开发,实现一体化监测点云数据的加工、对比分析及建模处理。通过结果验证显示：①改进后的算法效率高、精度优,适用于大坝点云数据处理和对比分析;②开发的点云数据一体化处理软件简单易用,能够精准反映坝体变形情况;③黄龙带坝体稳定,尚无安全隐患。