ISSN 1000-1239 CN 11-1777/TP

计算机研究与发展 ›› 2021, Vol. 58 ›› Issue (12): 2748-2759.doi: 10.7544/issn1000-1239.2021.20200595

• 人工智能 • 上一篇    下一篇

3D物体检测的异构方法

吕卓1,2,3,4,姚治成1,2,5,贾玉祥4,包云岗1,2,5   

  1. 1(中国科学院计算技术研究所 北京 100190);2(计算机体系结构国家重点实验室(中国科学院计算技术研究所) 北京 100190);3(数学工程与先进计算国家重点实验室 郑州 450001);4(郑州大学信息工程学院 郑州 450001);5(中国科学院大学 北京 100049) (lvzhuo11@163.com)
  • 出版日期: 2021-12-01
  • 基金资助: 
    军科委基础加强项目(2019-xCxQ-xD-172-00);广东省普及型高性能计算机重点实验室项目(2017B030314073);国家自然科学基金项目(62090020, 61672499);中国科学院青年促进创新会项目(2013073);中国科学院战略性先导科技专项(XDC05030200)

A Heterogeneous Approach for 3D Object Detection

Lü Zhuo1,2,3,4, Yao Zhicheng1,2,5, Jia Yuxiang4, Bao Yungang1,2,5   

  1. 1(Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190);2(State Key Laboratory of Computer Architecture (Institute of Computing Technology, Chinese Academy of Sciences), Beijing 100190);3(State Key Laboratory of Mathematical Engineering and Advanced Computing, Zhengzhou 450001);4(School of Information Engineering, Zhengzhou University, Zhengzhou 450001);5(University of Chinese Academy of Sciences, Beijing 100049)
  • Online: 2021-12-01
  • Supported by: 
    This work was supported by the Foundation Enhancement Project of Commission of Science and Technology of the CMC(2019-xCxQ-xD-172-00), the Guangdong Province Key Laboratory of Popular High Performance Computers (2017B030314073), the National Natural Science Foundation of China (62090020, 61672499), the Youth Innovation Promotion Association of Chinese Academy of Sciences (2013073), and the Strategic Priority Research Program of Chinese Academy of Sciences (XDC05030200).

摘要: 3D物体检测是计算机视觉的一个重要研究方向,在自动驾驶等领域有着广泛的应用.现有的前沿工作采用端到端的深度学习方法,虽然达到了很好的检测效果但存在着算法复杂度高、计算量大、实时性不够等问题.经过分析发现3D物体检测中的“部分任务”并不适合使用深度学习的方法进行解决,为此提出了一种基于异构方法的3D物体检测方法,该方法在检测过程中同时使用深度学习和传统算法,将检测过程划分为多任务阶段:1)利用深度学习方法从被检测图片中获取被检测物体的mask、物体类别等信息;2)基于mask,利用快速聚类方法从雷达点云空间中筛选出目标物体的表面雷达点;3)利用物体mask、类别、雷达点云等信息计算物体朝向、边框等信息,最终实现3D物体检测.对该方法进行了系统实现,称之为HA3D(a heterogeneous approach for 3D object detection).经实验表明:在针对汽车的3D检测数据集KITTI上,该方法与代表性的基于深度学习的3D物体检测方法相比,在检测精度下降接受范围内(2.0%),速度提升了52.2%,精确率与计算时间的比值提升了49%.从综合表现上来看,方法具有明显的优势.

关键词: 深度学习, 自动驾驶, 实例分割, 聚类, 3D物体检测

Abstract: 3D object detection is an important research direction of computer vision, and has a wide range of applications in areas such as autonomous driving. Existing cutting-edge works use end-to-end deep learning methods. Although it has achieved good detection results, it has problems such as high algorithm complexity, large calculation volume, and insufficient real-time performance. After analysis, we found that the deep learning method is not suitable for solving “partial tasks” in 3D object detection. For this reason, this paper proposes a 3D object detection scheme based on heterogeneous methods. This method uses both deep learning and traditional algorithms in the detection process, and divides the detection process into multi-task stages: 1)Use deep learning methods to obtain information such as the mask and object category of the detected object from the detected picture; 2) Based on the mask, use the fast clustering method to filter out the surface radar points of the target object from the radar point cloud space; 3) Use the information such as the object’s mask, category and radar point cloud to calculate the object’s orientation, border and other information to finally realize 3D object detection. We have implemented this method systematically, which we call HA3D (a heterogeneous approach for 3D object detection). Experiments show that on the 3D detection data set KITTI for cars, the method in this paper is within the acceptance range of detection accuracy decline (2.0%) compared with the representative 3D object detection method based on deep learning, the speed is increased by 52.2%. The ratio of the accuracy to the calculation time has increased by 49%. From the perspective of comprehensive performance, this method has obvious advantages.

Key words: deep learning, autonomous driving, instance segmentation, clustering, 3D object detection

中图分类号: