[1]蔡黎明,杜吉祥,刘怀进,等.动态卷积的3D点云目标检测算法[J].华侨大学学报(自然科学版),2023,44(1):111-118.[doi:10.11830/ISSN.1000-5013.202204030]
　CAI Liming,DU Jixiang,LIU Huaijin,et al.3D Point Cloud Target Detection Algorithm Based on Dynamic Convolution[J].Journal of Huaqiao University(Natural Science),2023,44(1):111-118.[doi:10.11830/ISSN.1000-5013.202204030]
点击复制

动态卷积的3D点云目标检测算法()
分享到：

《华侨大学学报（自然科学版）》[ISSN:1000-5013/CN:35-1079/N]

卷:
第44卷
期数:
2023年第1期
页码:
111-118
栏目:
出版日期:
2023-01-10

文章信息/Info

Title:
3D Point Cloud Target Detection Algorithm Based on Dynamic Convolution
文章编号:
1000-5013(2023)01-0111-08
作者:
蔡黎明1; 2; 3;  杜吉祥1; 2; 3;  刘怀进1; 2; 3;  张洪博1; 2; 3;  黄敬东1; 2; 3
1. 华侨大学 计算机科学与技术学院, 福建 厦门 361021;2. 华侨大学 福建省大数据智能与安全重点实验室, 福建 厦门 361021;3. 华侨大学 厦门市计算机视觉与模式识别重点实验室, 福建 厦门 361021
Author(s):
CAI Liming1; 2; 3;  DU Jixiang1; 2; 3;  LIU Huaijin1; 2; 3;  ZHANG Hongbo1; 2; 3;  HUANG Jingdong1; 2; 3
1. College of Computer Science and Technology, Huaqiao University, Xiamen 361021, China; 2. Fujian Key Laboratory of Big Data Intelligence and Security, Huaqiao University, Xiamen 361021, China; 3. Xiamen Key Laborstory of Computer Vision and Pattern Recognition, Huaqiao University, Xiamen 361021, China
关键词:
点云;  3D目标检测;  动态卷积;  分类回归
Keywords:
point cloud;  3D target detection;  dynamic convolution;  classification regression
分类号:
TP391.41
DOI:
10.11830/ISSN.1000-5013.202204030
文献标志码:
A
摘要:
针对不规则且稀疏的点的提取特征问题,提出一种以动态卷积作为特征提取的3D点云目标检测算法.首先,以一种新型的动态卷积的方式自适应学习点的位置特征,分类出前景点与背景点,同时对提取出的前景点逐一做回归框;然后,用非极大值抑制选出分数值最好的回归框.其次,进行粒度的细化,得到修正规范的3D回归框,完成3D物体的目标检测.最后,在KITTI数据集上验证算法的有效性.结果表明:文中所提算法在汽车类、行人类、自行车类数据集上的3D点云目标检测精度更高.
Abstract:
Aiming at the problem of extracting features from irregular and sparse points, a 3D point cloud target detection algorithm using dynamic convolution as feature extraction is proposed. Firstly, a new dynamic convolution method is used to adaptively learn the position features of points, and classify the foreground points and background points. At the same time, the extracted foreground points are used as regression boxes one by one. Then, non maximum suppression is used to select the regression box with the best score, secondly, the granularity is refined to obtain a 3D regression box of the revised standard, and the target detection of 3D objects is completed. Finally, the validity of the algorithm is verified on KITTI data set. The results show that the proposed algorithm has higher detection precision of 3D point cloud target on the car, pedestrian and bicycle data sets.

参考文献/References:

[1] ZHOU Yin,TUZEL O.Voxelnet: End-to-end learning for point cloud based 3D object detection[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition.Salt Lake City:IEEE Press,2018:4490-4499.
[2] YAN Yan,MAO Yuxing,LI Bo.Second: Sparsely embedded convolutional detection[J].Sensors,2018,18(10):3337.DOI:10.3390/s18103337.
[3] GRAHAM B,ENGELCKE M,VAN DER MAATEN L.3D semantic segmentation with submanifold sparse convolutional networks[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition.Salt Lake City:IEEE Press,2018:9224-9232.DOI:10.48550/arXiv.1711.10275.
[4] LANG A H,VORA S,CAESAR H,et al.PointPillars: Fast encoders for object detection from point clouds[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition.Long Beach:IEEE Press,2019:12689-12697.DOI:10.1109/CVPR.2019.01298.
[5] SHI Shaoshuai,GUO Chaoxu,JIANG Li,et al.PV-RCNN: Point-Voxel feature set abstraction for 3D object detection[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition.Long Beach:IEEE Press,2021:10529-10538.DOI:10.48550/arXiv.1912.13192.
[6] DENG Jiajun,SHI Shaoshuai,LI Peiwei,et al.Voxel R-CNN: Towards high performance Voxel-based 3D object detection[C]//AAAI Conference on Artificial Intelligence.[S.l.]:AAAI Press,2021:1201-1209.
[7] MA Ruiqi,CHEN Chi,YANG Bisheng,et al.CG-SSD: Corner guided single stage 3D object detection from LiDAR point cloud[EB/OL].(2022-02-24)[2022-03-28] .https://doi.org/10.48550/arXiv.2202.11868.
[8] XU Wencai,HU Jie,CHEN Ruinan,et al.Keypoint-aware single-stage 3D object detector for autonomous driving[J].Sensors,2022,22:1451.DOI:10.3390/s22041451.
[9] QI C R,SU Hao,MO Kaichun,et al.PointNet: Deep learning on point sets for 3D classification and segmentation[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition.Honolulu:IEEE Press,2017:77-85.DOI:10.1109/CVPR.2017.16.
[10] QI C R, YI Li, SU Hao, et al.PointNet++: Deep hierarchical feature learning on point sets in a metric space [C]//Annual Conference on Neural Information Processing Systems.Long Beach:NIPS Press,2017:5099-5108.
[11] SHI Shaoshuai,WANG Xiaogang,LI Hongsheng.PointRCNN: 3D object proposal generation and detection from point cloud[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition.Los Angeles:IEEE Press,2019:770-779.DOI:10.48550/arXiv.1812.04244.
[12] YANG Zetong,SUN Yanan,LIU Shu,et al.3DSSD: Point-based 3D single stage object detector[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition.Seattle:IEEE Press,2020:11040-11048.DOI:10.1109/CVPR42600.2020.01105.
[13] SHI Weijing,RAJKUMA R.Point-GNN: Graph neural network for 3D object detection in a point cloud[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition.Seattle:IEEE Press,2020:1708-1716.DOI:10.1109/CVPR42600.2020.00178.
[14] XU Mutian,DING Runyu,ZHAO Hengshuang,et al.PAConv: Position adaptive convolution with dynamic kernel assembling on point clouds[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition.[S.l.]:IEEE Press,2021:3172-3181.DOI:10.48550/arXiv.2103.14635.
[15] LIN T Y,GOYAL P,GIRSHICK R,et al.Focal loss for dense object detection[C]//IEEE International Conference on Computer Vision.Venice:IEEE Press,2017:2999-3007.DOI:10.1109/TPAMI.2018.2858826.
[16] CHEN Xiaozhi,MA Huimin,WAN Ji,et al.Multi-view 3D object detection network for autonomous driving[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition.Honolulu:IEEE Press,2017:6526-6534.DOI:10.1109/CVPR.2017.691.
[17] KU J,MOZIFIAN M,LEE J,et al.Joint 3D proposal generation and object detection from view aggregation[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems.Madrid:IEEE Press,2018:5750-5757.DOI:10.48550/arXiv.1712.02294.
[18] QI C R,LIU Wei,WU Chenxia,et al.Frustum PointNets for 3D object detection from RGB-D data[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition.Salt Lake City:IEEE Press,2018:918-927.DOI:10.48550/arXiv.1711.08488.

备注/Memo

备注/Memo:
收稿日期: 2022-04-26
通信作者: 杜吉祥(1977-),男,教授,博士,博士生导师,主要从事模式识别及图像处理的研究.E-mail:jxdu@hqu.edu.cn.
基金项目: 国家自然科学基金资助项目(61673186); 国家重点研发计划专项项目子课题(2019YFC1604705); 福建省厦门市科技计划项目(3502220193037)http://www.hdxb.hqu.edu.cn
更新日期/Last Update: 2023-01-20