基于改进PointNet++的城市道路点云分类方法

doi:10.16088/j.issn.1001-6600.2024060301

摘要/Abstract

摘要： 城市道路场景的点云数据量巨大、类别分布不平衡且密度极不均匀,导致现有的点云分类方法难以满足高精度分类的需求。为了解决现有PointNet++网络对局部特征提取不充分的问题,本文充分考虑场景的上下文信息和点之间的全局依赖性,构建融合上下文信息的PointNet++点云分类网络模型。首先,基于注意力机制设计局部特征聚合模块,通过动态地融合邻域点特征以充分捕获局部信息。其次,考虑现有的分类模型不能顾及上下文信息,导致复杂场景下的分类性能受限,本文构建上下文感知模块和双注意力模块,从多个维度提取上下文信息,进一步增强特征的表达能力。实验结果表明:改进模型在大型点云数据集下具有更高的分类精度及更强的泛化性能(总体分类精度在Oakland和Paris公开数据集上分别为98.70%和96.84%),更适用于大规模点云分类。

关键词: 点云分类, PointNet++, 局部特征, 注意力机制, 上下文信息, 城市道路

Abstract: The large amount of point cloud data, unbalanced distribution and uneven density of urban road scenes make it difficult for the current point cloud classification methods to meet the requirements of high-precision classification. To deal with the problem of insufficient local feature extraction by PointNet++ networks, a local feature aggregation module is designed based on the attention mechanism, which adequately captures local information by dynamically merging neighboring point features. Considering that the existing classification models cannot take into account contextual information, which leads to limited classification performance in complex scenes, a dual-attention module and a context-aware module are constructed to extract contextual information from several dimensions to further enhance the feature representation capability. The experimental results show that the new method has higher accuracy and stronger generalization performance (overall accuracy reaches 98.70% and 96.84% in Oakland and Paris publicly available datasets) under large point cloud datasets, and is more suitable for large-scale point cloud classification.

Key words: point cloud classification, PointNet++, local feature, attention mechanism, contextual information

中图分类号: U495;TP391.41;TP18

田晟, 熊辰崟, 龙安洋. 基于改进PointNet++的城市道路点云分类方法[J]. 广西师范大学学报（自然科学版）, 2025, 43(4): 1-14.

TIAN Sheng, XIONG Chenyin, LONG Anyang. Point Cloud Classification Method of Urban Roads Based on Improved PointNet++[J]. Journal of Guangxi Normal University(Natural Science Edition), 2025, 43(4): 1-14.

参考文献

[1] 许梦兵,韩友美,王留召,等.基于激光点云的高精度井盖提取和沉降病害检测方法研究[J].中国激光,2021,48(16): 1604001. DOI: 10.3788/CJL202148.1604001.
[2] 方莉娜,沈贵熙,游志龙,等.融合点云和多视图的车载激光点云路侧多目标识别[J].测绘学报,2021,50(11): 1558-1573.
[3] 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. DOI: 10.14132/j.cnki.1673-5439.2019.04.010.
[4] 钱志鸿,田春生,郭银景,等.智能网联交通系统的关键技术与发展[J].电子与信息学报,2020,42(1): 2-19. DOI: 10.11999/JEIT190787.
[5] ZHAO Z H, SONG Y Y, CUI F C, et al. Point cloud features-based kernel SVM for human-vehicle classification in millimeter wave radar[J]. IEEE Access, 2020, 8: 26012-26021. DOI: 10.1109/ACCESS.2020.2970533.
[6] 熊艳,高仁强,徐战亚.机载LiDAR点云数据降维与分类的随机森林方法[J].测绘学报,2018,47(4): 508-518. DOI: 10.11947/j.AGCS.2018.20170417.
[7] AGUILAR F J, NEMMAOUI A, AGUILAR M A, et al. Building tree allometry relationships based on TLS point clouds and machine learning regression[J]. Applied Sciences, 2021, 11(21): 10139. DOI: 10.3390/app112110139.
[8] 张毅,姚丹亚,李力,等.智能车路协同系统关键技术与应用[J].交通运输系统工程与信息,2021,21(5): 40-51. DOI: 10.16097/j.cnki.1009-6744.2021.05.005.
[9] SUN T B, LIU J H, KAN J M, et al. A study on the classification of vegetation point cloud based on random forest in the straw checkerboard barriers area[J]. Journal of Intelligent & Fuzzy Systems, 2021,41(3): 4337-4349. DOI: 10.3233/JIFS-189694.
[10] TANG Q Y, ZHANG L T, LAN G W, et al. A classification method of point clouds of transmission line corridor based on improved random forest and multi-scale features[J]. Sensors, 2023, 23(3): 1320. DOI: 10.3390/s23031320.
[11] 汤圣君,张韵婕,李晓明,等.超体素随机森林与LSTM神经网络联合优化的室内点云高精度分类方法[J].武汉大学学报(信息科学版),2023,48(4): 525-533. DOI: 10.13203/j.whugis20220125.
[12] 田晟,龙安洋.基于图卷积和多层特征融合的点云分类方法[J].激光与光电子学进展,2023,60(14):281-288.
[13] ZHAO N, CHUA T S, LEE G H. PS²-Net: a locally and globally aware network for point-based semantic segmentation[C]//2020 25th International Conference on Pattern Recognition (ICPR). Los Alamitos, CA: IEEE Computer Society, 2021: 723-730. DOI: 10.1109/ICPR48806.2021.9412875.
[14] 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). Los Alamitos, CA: IEEE Computer Society, 2017: 77-85. DOI: 10.1109/CVPR.2017.16.
[15] QI C R, YI L, SU H, et al. PointNet++: deep hierarchical feature learning on point sets in a metric space[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. Red Hook, NY: Curran Associates Inc., 2017: 5105-5114.
[16] WANG L, HUANG Y C, HOU Y L, et al. Graph attention convolution for point cloud semantic segmentation[C]// 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2019: 10288-10297. DOI: 10.1109/CVPR.2019.01054.
[17] CHEN C, FRAGONARA L Z, TSOURDOS A. GAPointNet: graph attention based point neural network for exploiting local feature of point cloud[J]. Neurocomputing, 2021,438: 122-132. DOI: 10.1016/j.neucom.2021.01.095.
[18] LI H C, XIONG P F, AN J, et al. Pyramid attention network for semantic segmentation[EB/OL]. (2018-11-25)[2024-06-01]. https://arxiv.org/abs/1805.10180. DOI: 10.48550/arXiv.1805.10180.
[19] 陈涵娟,达飞鹏,盖绍彦.基于竞争注意力融合的深度三维点云分类网络[J].浙江大学学报(工学版),2021,55(12): 2342-2351. DOI: 10.3785/j.issn.1008-973X.2021.12.014.
[20] CHEN L Z, LI X Y, FAN D P, et al. LSANet: feature learning on point sets by local spatial aware layer[EB/OL]. (2019-06-20)[2024-06-01]. https://arxiv.org/abs/1905.05442. DOI: 10.48550/arXiv.1905.05442.
[21] FU J, LIU J, TIAN H J, et al. Dual attention network for scene segmentation[C]// 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2019: 3141-3149. DOI: 10.1109/CVPR.2019.00326.
[22] XIE S N, LIU S N, CHEN Z Y, et al. Attentional ShapeContextNet for point cloud recognition[C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Los Alamitos, CA: IEEE Computer Society, 2018: 4606-4615. DOI: 10.1109/CVPR.2018.00484.
[23] FENG M T, ZHANG L, LIN X F, et al. Point attention network for semantic segmentation of 3D point clouds[J]. Pattern Recognition, 2020, 107: 107446. DOI: 10.1016/j.patcog.2020.107446.
[24] WANG Y, SUN Y B, LIU Z W, et al. Dynamic graph CNN for learning on point clouds[J]. ACM Transactions on Graphics, 2019, 38(5): 146. DOI: 10.1145/3326362.

相关文章 15

[1]	韩烁, 江林峰, 杨建斌. 基于注意力机制PINNs方法求解圣维南方程[J]. 广西师范大学学报（自然科学版）, 2025, 43(4): 58-68.
[2]	石天怡, 南新元, 郭翔羽, 赵濮, 蔡鑫. 基于改进ConvNeXt的苹果叶片病害分类算法[J]. 广西师范大学学报（自然科学版）, 2025, 43(4): 83-96.
[3]	卢展跃, 陈艳平, 杨卫哲, 黄瑞章, 秦永彬. 基于掩码注意力与多特征卷积网络的关系抽取方法[J]. 广西师范大学学报（自然科学版）, 2025, 43(3): 12-22.
[4]	郭翔羽, 石天怡, 陈燕楠, 南新元, 蔡鑫. 基于YOLO-CDBW模型的列车接触网异物检测研究[J]. 广西师范大学学报（自然科学版）, 2025, 43(2): 56-69.
[5]	苏春海, 夏海英. 抗噪声双约束网络的面部表情识别[J]. 广西师范大学学报（自然科学版）, 2025, 43(2): 70-82.
[6]	刘玉娜, 马双宝. 基于改进YOLOv8n的轻量化织物疵点检测算法[J]. 广西师范大学学报（自然科学版）, 2025, 43(2): 83-94.
[7]	戴林华, 黎远松, 石睿, 何忠良, 李雷. HSED-YOLO:一种轻量化的带钢表面缺陷检测模型[J]. 广西师范大学学报（自然科学版）, 2025, 43(2): 95-106.
[8]	余快, 宋宝贵, 邵攀, 余翱. 基于层级尺度交互的U-Net遥感影像建筑物提取方法[J]. 广西师范大学学报（自然科学版）, 2025, 43(2): 121-132.
[9]	卢家辉, 陈庆锋, 王文广, 余谦, 何乃旭, 韩宗钊. 基于多尺度注意力的器官图像分割方法[J]. 广西师范大学学报（自然科学版）, 2024, 42(6): 138-148.
[10]	杜帅文, 靳婷. 基于用户行为特征的深度混合推荐算法[J]. 广西师范大学学报（自然科学版）, 2024, 42(5): 91-100.
[11]	田晟, 胡啸. 基于Transformer模型的车辆轨迹预测[J]. 广西师范大学学报（自然科学版）, 2024, 42(3): 47-58.
[12]	王天雨, 袁嘉伟, 齐芮, 李洋. 多类型知识增强的微博立场检测模型[J]. 广西师范大学学报（自然科学版）, 2024, 42(1): 79-90.
[13]	肖宇庭, 吕晓琪, 谷宇, 刘传强. 基于拆分残差网络的糖尿病视网膜病变分类[J]. 广西师范大学学报（自然科学版）, 2024, 42(1): 91-101.
[14]	席凌飞, 伊力哈木·亚尔买买提, 刘雅洁. 基于改进YOLOv5的铝型材表面缺陷检测方法[J]. 广西师范大学学报（自然科学版）, 2024, 42(1): 111-119.
[15]	宋冠武, 陈知明, 李建军. 基于ResNet-50的级联注意力遥感图像分类[J]. 广西师范大学学报（自然科学版）, 2023, 41(6): 80-91.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed