[1]陈义飞,郭胜,潘文安,等.基于多源传感器数据融合的三维场景重建[J].郑州大学学报(工学版),2021,42(2):81-87.[doi:10.13705/j.issn.1671-6833.2021.02.008]
　Chen Yifei,Guo Sheng,Pun Wan-On,et al.3D Scene Reconstruction Based on Multi-source Sensor Data Fusion[J].Journal of Zhengzhou University (Engineering Science),2021,42(2):81-87.[doi:10.13705/j.issn.1671-6833.2021.02.008]
点击复制

基于多源传感器数据融合的三维场景重建()
分享到：

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:
42
期数:
2021年2期
页码:
81-87
栏目:
出版日期:
2021-04-12

文章信息/Info

Title:
3D Scene Reconstruction Based on Multi-source Sensor Data Fusion
作者:
陈义飞1,郭胜2,潘文安2,陆彦辉1,3
1.郑州大学 信息工程学院，河南 郑州 450001；2.香港中文大学(深圳) 理工学院，广东 深圳 518172；3.深圳市大数据研究院，广东 深圳 518172
Author(s):
1.School of Information Engineering, Zhengzhou University, Zhengzhou 450001, China; 2.School of Science and Engineering, The Chinese University of Hong Kong, Shenzhen 518172, China; 3.Shenzhen Research Institute of Big Data, Shenzhen 518172, China
关键词:
Keywords:
data fusion;  3D modeling;  deep learning;  object detection;  feature matching;  scene recurrence
DOI:
10.13705/j.issn.1671-6833.2021.02.008
文献标志码:
A
摘要:
针对特定校园场景重建结果中存在的目标冗余等情况 ， 提出了一种相机 RGB 位图和激光雷达数据融合的方法 。 在三维重建领域 ， 通过数据融合的方法剔除特定场景中无关目标以实现三维场景重现 。 首先使用轻量级的 LeGO-LOAM 算法 ， 将不同类型的特征点进行特征提取与匹配 ， 融合不同时刻的点云完成点云地图的重现 ； 然后对构建的点云地图中可能存在的无关目标 ， 借助多源传感器数据和深度学习在计算机视觉领域中的应用技术 ， 在三维空间中进行目标检测与剔除 。 对于点云地图建模与目标检测 3 个不同过程 ， 采用点云配准的方法对其进行算法融合 ， 最终完成校园环境下的场景重现 。 实验结果表明 ， 基于多源数据融合的方法可有效地将三维建模与目标检测 2 个过程进行结合 ， 完成校园场景下无冗余目标的点云地图构建 。 该方法可应用于智慧城市 、 无人驾驶等领域 ， 具有实际应用价值 。
Abstract:
Aiming at the target redundancy in reconstruction of certain scenes, in this paper a data fusion method of the camera RGB bitmap and lidar data was employed to solve the problem. In the field of 3D reconstruction, this method of data fusion could eliminate the irrelevant targets in the specific scene and reproduce the three-dimensional scene. The lightweight LeGO-LOAM algorithm was used to extract and match different types of feature points at first, and point clouds were merged at different times to complete the reproduction of the point cloud map. For the irrelevant targets in the constructed point cloud map, with the help of multi-source sensor data and deep learning application technology in the field of computer vision, object detection and elimination are accomplished in three-dimensional space. For the two different processes of point cloud modeling and target detection, the method of point cloud registration was adopted to fuse the algorithm and finally complete the scene reproduction in the campus environment. Experimental results showed that the method based muti-source data fusion could effectively combine the two processes of 3D modeling and object detection. The method proposed in this paper could be applied to smart cities, unmanned driving and other fields, and should have practical application value.

参考文献/References:

[1] 张勤, 贾庆轩. 基于激光与单目视觉的室内场景三维重建[J]. 系统仿真学报, 2014, 26(2):357-362.

[2] 辛俊伟, 罗艳, 程朋根. 基于LiDAR技术的古建筑复杂曲面三维重建[J]. 测绘与空间地理信息, 2017, 40(7):65-67.
[3] 刘晋,禹鹏,白隽瑄,等.基于三维激光扫描技术的犯罪现场重建[J].刑事技术,2017,42(6):476-481.
[4] 李云翔.相机标定与三维重建技术研究[D].青岛:青岛大学,2009.
[5] 曹明玮,钱烨强,王冰,等.基于监督式学习的全景相机与激光雷达的联合标定[J].机电一体化,2018,24(1):3-9,34.
[6] ZHANG J,SINGH S.Low-drift and real-time lidar odometry and mapping[J].Autonomous robots,2017,41(2):401-416.
[7] DISSANAYAKE M W M G,NEWMAN P,CLARK S,et al.A solution to the simultaneous localization and map building (SLAM) problem[J].IEEE transactions on robotics and automation,2001,17(3):229-241.
[8] SHAN T X,ENGLOT B.LeGO-LOAM:lightweight and ground-optimized lidar odometry and mapping on variable terrain[C]//2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).New York:IEEE,2018:4758-4765.
[9] CHARLES R Q,HAO S,MO K C,et al.PointNet:deep learning on point sets for 3D classification and segmentation[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).New York:IEEE,2017:77-85.
[10] QI C R,YI L,SU H,et al.PointNet++:deep hierarchical feature learning on point sets in a metric space[EB/OL].(2017-06-07)[2019-12-01].https://arxiv.org/abs/1706.02413.
[11] 张艺琨,唐雁,陈强.基于多特征融合的三维模型检索[J].郑州大学学报(工学版),2019,40(1):1-6.
[12] LI Y Y, BU R, SUN M C, et al. PointCNN: convolution on x-transformed points[C]//32nd Conference on Neural Information Processing Systems. New York:ACM，2018:820-830.
[13] QI C R,LIU W,WU C X,et al.Frustum PointNets for 3D object detection from RGB-D data[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.New York:IEEE,2018:918-927.
[14] 庄加福.基于机载激光雷达的复杂场景车辆类目标检测[D].武汉:华中科技大学,2013.
[15] 叶语同,李必军,付黎明.智能驾驶中点云目标快速检测与跟踪[J].武汉大学学报(信息科学版),2019,44(1):139-144,152.
[16] GEIGER A,LENZ P,STILLER C,et al.Vision meets robotics:the KITTI dataset[J].The international journal of robotics research,2013,32(11):1231-1237.[17] LIU Y S,HU Z C,GAO S,et al.In situ analysis of major and trace elements of anhydrous minerals by LA-ICP-MS without applying an internal standard[J].Chemical geology,2008,257(1/2):34-43.
[18] RUSINKIEWICZ S,LEVOY M.Efficient variants of the ICP algorithm[C]//Proceedings 3rd International Conference on 3-D Digital Imaging and Modeling.New York:IEEE,2001:145-152.
[19] 戴静兰,陈志杨,叶修梓.ICP算法在点云配准中的应用[J].中国图象图形学报,2007,12(3):517-521.

更新日期/Last Update: 2021-05-30