欢迎访问《图学学报》 分享到：
图学学报

图学学报

• 专论：第12届中国计算机图形学大会 (CHINAGRAPH 广州) • 上一篇    下一篇

基于点云数据的三维目标识别和模型分割方法

  

  1. 1. 中国石油大学(华东)计算机与通信工程学院，山东 青岛 266580； 
    2. 中国科学院计算技术研究所智能信息处理重点实验室，北京 100190； 
    3. 中国科学院大学，北京 100190
  • 出版日期:2019-04-30 发布日期:2019-05-10
  • 基金资助:
    中央高校基本科研业务费专项资金项目(18CX06049A)；国家自然科学基金项目(61379106，61379082，61227802)；山东省自然科学基金 项目(ZR2015FM011，ZR2013FM036)

3D Object Recognition and Model Segmentation Based on Point Cloud Data

  1. 1. College of Computer and Communication Engineering, China University of Petroleum, Qingdao Shandong 266580, China; 
    2. Key Laboratory of Intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190, China; 
    3. University of Chinese Academy of Sciences, Beijing 100190, China
  • Online:2019-04-30 Published:2019-05-10

PDF (PC)

298

摘要： 三维模型的深度特征表示是三维目标识别和三维模型语义分割的关键和前提，在 机器人、自动驾驶、虚拟现实、遥感测绘等领域有着广泛的应用前景。然而传统的卷积神经网 络需要以规则化的数据作为输入，对于点云数据需要转换为视图或体素网格来处理，过程复杂 且损失了三维模型的几何结构信息。借助已有的可以直接处理点云数据的深度网络，针对产生 的特征缺少局部拓扑信息问题进行改进，提出一种利用双对称函数和空间转换网络获得更鲁棒、 鉴别力更强的特征。实验表明，通过端到端的方式很好地解决缺少局部信息问题，在三维目标 识别、三维场景语义分割任务上取得了更好的实验效果，并且相比于 PointNet++在相同精度的 情况下训练时间减少了 20%。

关键词: 点云, 深度学习, 原始数据, 三维目标识别, 三维模型分割

Abstract: Deep representation of 3D model is the key and prerequisite for 3D object recognition and 3D model semantic segmentation, providing a wide range of applications ranging from robotics, automatic driving, virtual reality, to remote sensing and other fields. However, convolutional architectures require highly regular input data formats and most researchers typically transform point cloud data to regular 3D voxel grids or sets of images before feeding them to a deep net architecture. The process is complex and the 3D geometric structure information will be lost. In this paper, we make full use of the existing deep network which can deal with point cloud data directly, and propose a new algorithm that uses double symmetry function and space transformation network to obtain more robust and discriminating features. The local topology information is also incorporated into the final features. Experiments show that the proposed method solves the problem of lacking local information in an end-to-end way and achieves ideal results in the task of 3D object recognition and 3D scene semantic segmentation. Meanwhile, the method can save 20% training time compared to PointNet++with the same precision.

Key words: point cloud, deep learning, raw data, 3D object recognition, 3D model segmentation