基于双目点云的三维物体检测准确性分析

doi:10.7523/j.ucas.2020.0042

中国科学院大学学报 ›› 2022, Vol. 39 ›› Issue (5): 677-683.DOI: 10.7523/j.ucas.2020.0042

• 电子信息与计算机科学 • 上一篇下一篇

基于双目点云的三维物体检测准确性分析

刘王超^1,2,3, 娄鑫^1,2

1. 上海科技大学信息科学与技术学院, 上海 201210;
2. 中国科学院上海微系统与信息技术研究所, 上海 200050;
3. 中国科学院大学电子电气与通信工程学院, 北京 100049

收稿日期:2020-05-18 修回日期:2020-08-28 发布日期:2021-05-31
通讯作者: 娄鑫
基金资助:
国家自然科学基金（61801292）和上海青年科技英才扬帆计划（18YF1416600）资助

Accuracy analysis of 3D object detection based on stereo point cloud

LIU Wangchao^1,2,3, LOU Xin^1,2

1. School of Information Science & Technology, ShanghaiTech University, Shanghai 201220, China;
2. Shanghai Institute of Microsystem & Information Technology, Chinese Academy of Sciences, Shanghai 200050, China;
3. School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Science, Beijing 100049, China

Received:2020-05-18 Revised:2020-08-28 Published:2021-05-31

摘要/Abstract

摘要： 双目相机虽然能通过算法生成密集的深度数据，但其在精度上与激光雷达生成的深度数据有着较大的差距，特别是在纹理不明显的区域。针对这种现象，尝试使用激光雷达的精确点云来排除由双目相机产生的pseudo-LiDAR数据中与之差别较大的部分点，然后将优化后的pseudo-LiDAR用以进行三维物体检测。实验结果显示，将pseudo-LiDAR数据中的不准确点（坏点）排除，有助于提高检测准确率，最多可提高21.02%。因此，如何不依赖激光雷达数据来排除pseudo-LiDAR点云中的坏点是进一步提高双目相机系统检测准确率的关键。

关键词: 三维物体检测, 激光雷达, 双目相机, 点云, 误差

Abstract: 3D object detection is a crucial task in autonomous driving. Recently, the accuracy of LiDAR-based 3D object detection algorithms have improved dramatically. However, if we replace the LiDAR data with the depth map that generates from stereo cameras, the detection performance drops a lot. In a recent research, it is found that by transforming the stereo-based depth map to point cloud representation, referred to as pseudo-LiDAR, the performance of 3D object detection can be improved significantly. However, there is still a big gap of accuracy between LiDAR-based and stereo-based algorithms. One of the main reasons is that there are error points in the transformed pseudo-LiDAR data. To study this, we use the LiDAR points cloud to correct the pseudo-LiDAR data, i.e., to detect and exclude the error points. Then we use the optimized pseudo-LiDAR to perform 3D object detection. Experimental results show that the detection accuracy can be improved by as much as 21.02%.

Key words: 3D object detection, LiDAR, stereo cameras, point cloud, error

中图分类号:

TN47

刘王超, 娄鑫. 基于双目点云的三维物体检测准确性分析[J]. 中国科学院大学学报, 2022, 39(5): 677-683.

LIU Wangchao, LOU Xin. Accuracy analysis of 3D object detection based on stereo point cloud[J]. Journal of University of Chinese Academy of Sciences, 2022, 39(5): 677-683.

参考文献

[1] Lei Z Y, Wang Y, Xu Y F, et al. From coarse to fine:a monocular depth estimation model based on left-right consistency[C]//2019 IEEE 19th International Conference on Communication Technology (ICCT). October 16-19, 2019, Xi'an, China. IEEE, 2019:1621-1625.DOI:10.1109/ICCT46805.2019.8947220.
[2] Mayer N, Ilg E, Häuusser P, et al. A large dataset to train convolutional networks for disparity, optical flow, and scene flow estimation[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). June 27-30, 2016, Las Vegas, NV,USA. IEEE, 2016:4040-4048.DOI:10.1109/CVPR.2016.438.
[3] Chen X Z, Kundu K, Zhu Y K, et al. 3D object proposals using stereo imagery for accurate object class detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018, 40(5):1259-1272.DOI:10.1109/TPAMI.2017.2706685.
[4] Xu B, Chen Z Z. Multi-level fusion based 3D object detection from monocular images[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 18-23, 2018, Salt Lake City, UT,USA. IEEE, 2018:2345-2353.DOI:10.1109/CVPR.2018.00249.
[5] Liang M, Yang B, Wang S L, et al. Deep continuous fusion for multi-sensor 3D object detection[C]//European Conference on Computer Vision (ECCV). Munich, Germany:2018:663-678.
[6] Wang Y, Chao W, Garg D, et al. Pseudo-LiDAR from visual depth estimation:bridging the gap in 3D object detection for autonomous driving[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 15-20, 2019, Long Beach, CA, USA. IEEE, 2019:8437-8445.DOI:10.1109/CVPR.2019.00864.
[7] Chen X, Ma H, Wan J, et al. Multi-view 3D object detection network for autonomous driving[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). July 21-26,2017, Honolulu, HI, USA. IEEE, 2017:6526-6534.DOI:10.1109/CVPR.2017.691.
[8] Zhou Y, Tuzel O. VoxelNet:end-to-end learning for point cloud based 3D object detection[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR). June 18-23, 2018, Salt Lake City, UT, USA. IEEE, 2018:4490-4499.DOI:10.1109/CVPR.2018.00472.
[9] Charles R Q, Su H, 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). July 21-26, 2017, Honolulu, HI, USA. IEEE, 2017:77-85.DOI:10.1109/CVPR.2017.16.
[10] Guo X Y, Yang K, Yang W K, et al. Group-wise correlation stereo network[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 15-20, 2019, Long Beach, CA, USA. IEEE, 2019:3268-3277.DOI:10.1109/CVPR.2019.00339.
[11] Wu Z Y, Wu X Y, Zhang X P, et al. Semantic stereo matching with pyramid cost volumes[C]//2019 IEEE/CVF International Conference on Computer Vision (ICCV). October 27-November 2, 2019, Seoul, Korea(South):IEEE, 2019:7483-7492.DOI:10.1109/ICCV.2019.00758.
[12] Chang J R, Chen Y S. Pyramid stereo matching network[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 18-23, 2018, Salt Lake City, UT, USA. IEEE, 2018:5410-5418.DOI:10.1109/CVPR.2018.00567.
[13] Geiger A, Lenz P, Urtasun R. Are we ready for autonomous driving? The KITTI vision benchmark suite[C]//2012 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). June 16-21, 2012, Providence, RI, USA. IEEE, 2012:3354-3361.DOI:10.1109/CVPR.2012.6248074.
[14] Ku J, Mozifian M, Lee J, et al. Joint 3D proposal generation and object detection from view aggregation[C]//2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). October 1-5, 2018, Madrid, Spain. IEEE, 2018:1-8.DOI:10.1109/IROS.2018.8594049.
[15] 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 (CVPR). June 18-23, 2018, Salt Lake City, UT, USA. IEEE, 2018:918-927. DOI:10.1109/CVPR.2018.00102.

基于双目点云的三维物体检测准确性分析

Accuracy analysis of 3D object detection based on stereo point cloud

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

访问统计

联系我们

[1]	马曌月, 肖俊, 王颖. 基于面提取的三维岩体点云孔洞检测与修复方法[J]. 中国科学院大学学报, 2022, 39(5): 668-676.
[2]	刘丽娟, 王成, 聂胜, 朱笑笑, 习晓环, 王金亮. GEDI L2A不同算法对地面高程和森林冠层高度精度的影响分析[J]. 中国科学院大学学报, 2022, 39(4): 502-511.
[3]	张韬, 肖俊, 王颖. 基于匹配点逐层过滤的岩体点云配准方法[J]. 中国科学院大学学报, 2022, 39(3): 352-359.
[4]	梁磊, 习晓环, 王成, 王金亮, 黄友菊. 基于B/S架构的激光雷达电力巡线可视化管理与分析系统[J]. 中国科学院大学学报, 2022, 39(2): 201-207.
[5]	向建冰, 吕孝雷, 付希凯. 基于双曲等效的双星滑动聚束SAR两步成像算法[J]. 中国科学院大学学报, 2021, 38(4): 511-518.
[6]	黄芳, 甘淼, 于洋, 他志杰, 张海燕, 皮原月, 孙凌霄, 于瑞德. 基于CMIP5模式的2006—2017年中亚降水预估误差分析[J]. 中国科学院大学学报, 2021, 38(3): 333-340.
[7]	张苏, 齐立忠, 韩文军, 朱笑笑, 习晓环, 王成, 王濮, 聂胜. 基于无人机激光点云的树障检测与砍伐树木数量估算[J]. 中国科学院大学学报, 2020, 37(6): 760-766.
[8]	胡亮, 肖俊, 王颖. 基于多维度特征和MLP的岩体点云植被滤波方法[J]. 中国科学院大学学报, 2020, 37(3): 345-351.
[9]	王飞鹏, 肖俊, 王颖, 王云标. 一种基于高斯曲率的ICP改进算法[J]. 中国科学院大学学报, 2019, 36(5): 702-708.
[10]	王濮, 邢艳秋, 王成, 习晓环. 一种基于图割的机载LiDAR单木识别方法[J]. 中国科学院大学学报, 2019, 36(3): 385-391.
[11]	杜松, 李晓辉, 刘照言, 陈育伟, 王震, 唐伶俐, 贺文静. 激光雷达回波强度数据辐射特性分析[J]. 中国科学院大学学报, 2019, 36(3): 392-400.
[12]	赵海鹏, 习晓环, 王成, 雷钊. 基于车载激光扫描数据的城区道路自动提取[J]. 中国科学院大学学报, 2018, 35(6): 782-787.
[13]	尉驰俊, 邱越, 袁丁, 郑庭辉. 腹主动脉瘤形态学二维断面和三维立体测量的误差分析[J]. 中国科学院大学学报, 2018, 35(5): 607-611.
[14]	舒持恺, 候星甫, 王建金, 杨侃. 考虑随机观测误差影响的改进集对分析模型在水质模糊评价中的应用[J]. 中国科学院大学学报, 2018, 35(5): 627-634.
[15]	刘智凡, 王妙妙, 谢田法, 孙志华. 工具变量辅助的变系数测量误差模型的估计[J]. 中国科学院大学学报, 2018, 35(1): 1-9.