摘 要：针对点云的三维模型识别方法缺乏局部空间特征，从而影响 3D 模型的类识别的问题，提出一种基于残差模块的卷积神经网络三维模型识别方法。通过引入残差模块，构建深层神经网络增强点云模型的局部信息，提高物体的识别精度。同时，采用了一种获取多尺度局部空间信息的策略，加快了模型的推理能力。实验证明，算法识别准确率达到了 91.5%，加快了模型的推理速度，可应用于对点云模型识别有实时性要求的场景，如：流水线上物体的检测等。
关键词：三维模型识别；卷积神经网络；实时性
DOI:10.19850/j.cnki.2096-4706.2023.07.024
中图分类号：TP391.4 文献标识码：A 文章编号：2096-4706（2023）07-0093-05
Method of 3D Point Cloud Model Recognition Based on Neural Network
WU Pengcheng, LI Hui, WANG Fengcong
(Shenyang University of Technology, Shenyang 110870, China)
Abstract: Aiming at the problem that the 3D model recognition method of point cloud lacks local spatial features, which affects the class recognition of 3D model, a convolution neural network 3D model recognition method based on residual module is proposed. By introducing the residual module, a deep neural network is constructed to enhance the local information of the point cloud model and improve the object recognition accuracy. At the same time, a strategy of acquiring multi-scale local spatial information is adopted to accelerate the reasoning ability of the model. The experimental results prove that the recognition accuracy of the algorithm reaches 91.5%, which speeds up the reasoning speed of the model. And it can be applied to scenes that require real-time point cloud model recognition, such as object detection on the pipeline.
Keywords: 3D model recognition; convolution neural network; real time
参考文献：
[1] SUN J，OVSJANIKOV M，GUIBAS L. A Concise and Provably Informative Multi-Scale Signature Based on Heat Diffusion [J]. Computer Graphics Forum，2009，28（5）：1383-1392.
[2] OSADA R，FUNKHOUSER T，CHAZELLE B，et al. Shape Distributions [J].ACM Transactions on Graphics，2002，21（4）：807-832.
[3] SU H，MAJI S，KALOGERAKIS E，et al. Multi-View Convolutional Neural Networks for 3D Shape Recognition [J/OL].arXiv： 1505.00880 [cs.CV].（2015-05-05）.https://arxiv.org/abs/1505.00880.
[4] FENG Y F，ZHANG Z Z，ZHAO X B，et al. GVCNN：
Group-View Convolutional Neural Networks for 3D Shape Recognition [C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Salt Lake：IEEE，2018：264-272.
[5] MATURANA D，SCHERER S. VoxNet：A 3D Convolutional Neural Network for Real-Time Object Recognition [C]//2015 IEEE/RSJ International Conference on Intelligent Robots and Systems（IROS）.Hamburg：IEEE，2015：922-928.
[6] 杨军，王亦民.基于深度卷积神经网络的三维模型识别 [J].重庆邮电大学学报：自然科学版，2019，31（2）：253-260.
[7] 杨军，王顺，周鹏 . 基于深度体素卷积神经网络的三维模型识别分类 [J]. 光学学报，2019，39（4）：314-324.
[8] KLOKOV R，LEMPITSKY V. Escape from Cells：Deep Kd-Networks for the Recognition of 3D Point Cloud Models [C]//2017 IEEE International Conference on Computer Vision（ICCV）.Venice： IEEE，2017：863-872.
[9] BRONSTEIN M M，BRUNA J，LECUN Y，et al. Geometric Deep Learning：Going beyond Euclidean Data [J].IEEE Signal Processing Magazine，2017，34（4）：18-42.
[10] WANG Y，SUN Y B，LIU Z W，et al. Dynamic graph CNN for learning on Point Clouds [J/OL].arXiv:1801.07829 [cs.CV].（2018-01-24）.https://arxiv.org/abs/1801.07829.
[11] ZHANG K G，HAO M，WANG J，et al. Linked Dynamic Graph CNN：Learning on Point Cloud Via Linking Hierarchical Features [J/OL].arXiv:1904.10014 [cs.CV].（2019-04-22）.https://arxiv. org/abs/1904.10014.
[12] CHEN C，FRAGONARA L Z，TSOURDOS A. GAPNet： Graph Attention Based Point Neural Network for Exploiting Local Feature of Point Cloud [J/OL].arXiv：1905.08705 [cs.CV].（2019-05-21）.https://arxiv.org/abs/1905.08705.
[13] LIU X H，HAN Z Z，LIU Y-S，et al. Point2Sequence： Learning the Shape Representation of 3D Point Clouds with an AttentionBased Sequence to Sequence Network [J/OL].arXiv:1811.02565 [cs.CV]. （2019-05-21）.https://arxiv.org/abs/1811.02565v2.
[14] LI Y Y，BU R，SUN M C，et al. PointCNN：Convolution on X-Transformed Points [C]//Neural Information Processing Systems. Montréal：Curran Associates，2018：820-830.
[15] XU Y F，FAN T Q，XU M Y，et al. SpiderCNN：Deep Learning on Point Sets with Parameterized Convolutional Filters [J/ OL].arXiv:1803.11527 [cs.CV].（2018-03-30）.https://arxiv.org/ abs/1803.11527v3.
[16] ATZMON M，MARON H，LIPMAN Y. Point Convolutional Neural Networks by Extension Operators [J].ACM Transactions on Graphics （TOG），2018，37（4）：1-12.
[17] QI C R，SU H，KAICHUN M，et al. PointNet：Deep Learning on Point Sets for 3D Classification and Segmentation [C]//2017 IEEE Conference on Computer Vision and Pattern Recognition（CVPR）. Honolulu：IEEE，2017：77-85.
[18] QI C R，YI L，SU H，et al. PointNet++：Deep Hierarchical Feature Learning on Point Sets in a Metric Space [J/OL].arXiv:1706.02413 [cs.CV].（2017-06-07）.https://arxiv.org/abs/1706.02413.
[19] KLOKOV R，LEMPITSKY V. Escape from Cells：Deep Kd-Networks for the Recognition of 3D Point Cloud Models [C]//2017 IEEE International Conference on Computer Vision （ICCV）.Venice： IEEE，2017：863-872.
[20] WU Z R，SONG S R，KHOSLA A，et al. 3D ShapeNets： A Deep Representation for Volumetric Shapes [C]//2015 IEEE Conference on Computer Vision and Pattern Recognition （CVPR）. Boston：IEEE，2015：1912-1920.
作者简介：吴鹏程（1994.11—），男，汉族，四川广安人，硕士研究生在读，研究方向：基于点云模型的识别方法研究；李晖（1968.09—），女，汉族，山东蓬莱人，教授，博士，研究方向：网络通信与信号处理、信息安全、自然语言处理；王凤聪（1997.02—），男，汉族，山东聊城人，硕士研究生在读，研究方向：基于图像的目标检测方法研究。