首页 > 论文 > 激光与光电子学进展 > 57卷 > 20期(pp:201508--1)

基于改进Frustum PointNet的3D目标检测

3D Object Detection Based on Improved Frustum PointNet

  • 摘要
  • 论文信息
  • 参考文献
  • 被引情况
  • PDF全文
分享:

摘要

提出对图像和激光雷达点云数据进行3D目标检测的改进F-PointNet(Frustum PointNet)。首先利用图像的2D目标检测模型提取目标2D区域,并将其映射到点云数据中,得到该目标的点云候选区域,然后预测候选区域的3D目标掩模,最后利用掩模对3D目标进行检测。当预测掩模时,提出的宽阈值掩模处理可以用来减少原始网络的信息损失;增加注意力机制可以获取需要被关注的点和通道层;使用Focal Loss可以解决目标与背景不平衡的问题。通过多次对比实验,证明宽阈值掩模处理可以提高3D目标检测的准确率,同时注意力机制和Focal Loss可以提高预测的准确率。

Abstract

An improved F-PointNet (Frustum PointNet) for 3D target detection on image and lidar point cloud data is proposed. First, the 2D target detection model of the image is used to extract 2D region of the target, and it is mapped to the point cloud data to obtain the candidate region of the target. Then, the 3D target mask of the candidate region is predicted. Finally, the 3D target is detected by using mask. When the mask is predicted, the proposed wide-threshold mask processing is used to reduce the information loss of the original network, the attention mechanism is added to obtain the points and channel layers that require attention, the Focal Loss can solve the imbalance between the target and the background problem. Through multiple comparison experiments, it is proved that wide-threshold mask processing can improve the accuracy of 3D target detection, and the attention mechanism and Focal Loss can improve the accuracy of prediction.

广告组1 - 空间光调制器+DMD
补充资料

中图分类号:TN958.98

DOI:10.3788/LOP57.201508

所属栏目:机器视觉

基金项目:上海市科委基础研究项目;

收稿日期:2019-12-24

修改稿日期:2020-03-09

网络出版日期:2020-10-01

作者单位    点击查看

刘训华:东华大学信息科学与技术学院, 上海 201620东华大学数字化纺织服装技术教育部工程研究中心, 上海 201620
孙韶媛:东华大学信息科学与技术学院, 上海 201620东华大学数字化纺织服装技术教育部工程研究中心, 上海 201620
顾立鹏:东华大学信息科学与技术学院, 上海 201620东华大学数字化纺织服装技术教育部工程研究中心, 上海 201620
李想:东华大学信息科学与技术学院, 上海 201620东华大学数字化纺织服装技术教育部工程研究中心, 上海 201620

联系人作者:刘训华(XunHua_LIU@163.com)

备注:上海市科委基础研究项目;

【1】Zhang Y, Ren G Q, Cheng Z Y, et al. Application research of there-dimensional LiDAR in unmanned vehicle environment perception [J]. Laser & Optoelectronics Progress. 2019, 56(13): 130001.
张银, 任国全, 程子阳, 等. 三维激光雷达在无人车环境感知中的应用研究 [J]. 激光与光电子学进展. 2019, 56(13): 130001.

【2】Liu K B, Yang X H, He T T, et al. Indium phosphide-based near-infrared single photon avalanche photodiode detector arrays [J]. Laser & Optoelectronics Progress. 2019, 56(22): 220001.
刘凯宝, 杨晓红, 何婷婷, 等. InP基近红外单光子雪崩光电探测器阵列 [J]. 激光与光电子学进展. 2019, 56(22): 220001.

【3】Ji Y M, Chen Z Y, Tian P H, et al. A survey of 3D target detection methods in unmanned driving [J]. Journal of Nanjing University of Posts and Telecommunications (Natural Science Edition). 2019, 39(4): 72-79.
季一木, 陈治宇, 田鹏浩, 等. 无人驾驶中3D目标检测方法研究综述 [J]. 南京邮电大学学报(自然科学版). 2019, 39(4): 72-79.
Ji Y M, Chen Z Y, Tian P H, et al. A survey of 3D target detection methods in unmanned driving [J]. Journal of Nanjing University of Posts and Telecommunications (Natural Science Edition). 2019, 39(4): 72-79.
季一木, 陈治宇, 田鹏浩, 等. 无人驾驶中3D目标检测方法研究综述 [J]. 南京邮电大学学报(自然科学版). 2019, 39(4): 72-79.

【4】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), July 21-26, 2017, Honolulu, HI, USA. New York: IEEE. 2017, 6526-6534.

【5】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, June 18-23, 2018, Salt Lake City, UT, USA. New York: IEEE. 2018, 918-927.

【6】He K M, Gkioxari G, Dollar P, et al. Mask R-CNN [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2020, 42(2): 386-397.

【7】Qi C R, Yi L, Su H, et al. -07-07)[2019-12-23] . https:∥arxiv. 2017, org/abs/1706: 02413.

【8】Wan P. Object detection of 3D point clouds based on F-PointNet [J]. Journal of Shandong University (Engineering Science). 2019, 49(5): 98-104.
万鹏. 基于F-PointNet的3D点云数据目标检测 [J]. 山东大学学报(工学版). 2019, 49(5): 98-104.

【9】Woo S, Park J, Lee J Y, et al. CBAM: convolutional block attention module [M]. ∥Ferrari V, Hebert M, Sminchisescu C, et al. Computer vision-ECCV 2018. Lecture notes in computer science. Cham: Springer. 2018, 11211: 3-19.

【10】Lin T Y, Goyal P, Girshick R, et al. Focal loss for dense object detection [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2020, 42(2): 318-327.

【11】Lin T Y, Dollár P, Girshick R, et al. Feature pyramid networks for object detection . [C]∥2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), July 21-26, 2017, Honolulu, HI, USA. New York: IEEE. 2017, 936-944.

【12】Jaderberg M, Simonyan K, Zisserman A, et al. -02-04)[2019-12-23] . https:∥arxiv. 2016, org/abs/1506: 02025.

【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, June 16-21, 2012, Providence, RI, USA. New York: IEEE. 2012, 3354-3361.

【14】Liang M, Yang B, Wang S L, et al. Deep continuous fusion for multi-sensor 3D object detection [M]. ∥Ferrari V, Hebert M, Sminchisescu C, et al. Computer vision-ECCV 2018. Lecture notes in computer science. Cham: Springer. 2018, 11220: 663-678.

【15】Deng J, Czarnecki K. MLOD: a multi-view 3D object detection based on robust feature fusion method . [C]∥2019 IEEE Intelligent Transportation Systems Conference (ITSC), October 27-30, 2019, Auckland, New Zealand. New York: IEEE. 2019, 279-284.

引用该论文

Liu Xunhua,Sun Shaoyuan,Gu Lipeng,Li Xiang. 3D Object Detection Based on Improved Frustum PointNet[J]. Laser & Optoelectronics Progress, 2020, 57(20): 201508

刘训华,孙韶媛,顾立鹏,李想. 基于改进Frustum PointNet的3D目标检测[J]. 激光与光电子学进展, 2020, 57(20): 201508

您的浏览器不支持PDF插件,请使用最新的(Chrome/Fire Fox等)浏览器.或者您还可以点击此处下载该论文PDF