| 838 | 11 | 156 |
| 下载次数 | 被引频次 | 阅读次数 |
3D目标检测结合了深度信息,能够提供目标的位置、方向和大小等空间场景信息,在自动驾驶和机器人领域发展迅速。针对PV-RCNN在3D目标检测时不能够充分适应不同的物体尺度、不同的点云密度、部分变形和杂波等问题,对3D目标检测的任务进行实验研究。通过加入自适应可变形卷积、上下文融合模块和Gumbel Subset Sampling模块来训练层级特征,使得编码关键点自适应地朝着最具有判别和代表性的特征对齐,提高提案框回归精度。实验结果表明,改进后的PV-RCNN 3D目标检测精度得到了提升,尤其是在远距离物体识别和检测方面。
Abstract：3D target detection combines depth information to provide spatial scene information such as the position,direction, and size of the target, and is developing rapidly in the field of autonomous driving and robotics. Aiming at the problems that PV-RCNN can not fully adapt to different object scales, different point cloud densities, partial deformations and clutter in 3D target detection, experimental research on the task of 3D target detection is carried out. By adding adaptive deformable convolution, context fusion module and Gumbel Subset Sampling module to train hierarchical features, the coding key points are adaptively aligned to the most discriminative and representative features, and the regression accuracy of proposal box regression is improved. The experimental results show that the improved PV-RCNN 3D target detection accuracy has been improved, especially in long-distance object recognition and detection.
[1]CHEN X Z,MA H M,WAN J,et al.Multi-view 3D object detection network for autonomous driving[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Honolulu:IEEE,2017.
[2]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.
[3]YAN Y,MAO YX,LI B.SECOND:Sparsely embedded convolutional detection[J].Sensors,2018,18(10):1-17.
[4]黄莉芝.基于深度学习的目标检测算法研究[D].成都:西南交通大学,2018.
[5]HARA K,KATAOKA H,SATOH Y.Can spatiotemporal 3DCNNs retrace the history of 2DCNNs and Image Net?[C]//2018IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).Salt Lake:IEEE,2018.
[6]WANG DZ,POSNER I.Voting for voting in online point cloud object detection[C]//Robotics:Science and Systems.Rome,2015.
[7]L XM,DOU Q,CHEN H,et al.3D multi-scale FCN with random modality voxel dropout learning for Intervertebral Disc Localization and Segmentation from Multi-modality MR Images[J].Medical Image Analysis,2018,45:41-54.
[8]YU Y K,LI F,ZHENG Y,et al.Clustering-enhanced Point CNNfor point cloud classification learning[C]//2019 International Joint Conference on Neural Networks(IJCNN).Budapest,2019.
[9]SHI S,GUO C,JIANG L,et al.PV-RCNN:Point-voxel feature set abstraction for 3D object detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Seattle,2020:10526-10535.
[10]QI CR,SU H,MO K,et al.Pointnet:Deep learning on point sets for 3D classification and segmentation[C]//Proceedings of the IEEE conference on computer vision and pattern recognition.Honolulu,2017,77-85.
[11]GEIGER A,LENZ P,URTASUN R.Are we ready for autonomous driving?The KITTI vision benchmark suite[C]//IEEE Conference on Computer Vision&Pattern Recognition.Providence,2012.
[12]SUN P,KRETZSCHMAR H,DOTIWALLA X,et al.Scalability in perception for autonomous driving:Waymo open dataset[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).Seattle,2020.
[13]CAESAR H,BANKITI V,LANG A H,et al.nuscenes:Amultimodal dataset for autonomous driving[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition.Seattle,2020,11618-11628.
[14]张宏斌.基于点云的三维目标检测算法研究与标注工具设计[D].哈尔滨:哈尔滨工业大学,2019.
[15]闫岩.基于深度学习的点云三维目标检测算法研究[D].重庆:重庆大学,2019.
基本信息:
DOI：10.16791/j.cnki.sjg.2022.02.005
中图分类号:TP391.41;TP183
引用信息:
[1]傅荟璇,刘凌风,王宇超.基于改进PV-RCNN的3D目标检测算法实验研究[J].实验技术与管理,2022,39(02):23-28.DOI:10.16791/j.cnki.sjg.2022.02.005.
基金信息:
国家自然科学基金项目(52071112); 黑龙江省科学基金项目(F2017008); 中央高校基金项目(3072021CFJ0408)
2021-07-02
2021
2021-08-09
2021
1
2022-02-28
2022-02-28
2022-02-28