基于形状和骨架特征匹配的小样本三维点云目标识别方法

doi:10.12305/j.issn.1001-506X.2026.01.08

摘要/Abstract

摘要：

在非合作或遮蔽场景下，获取高质量目标三维点云具有困难性，小样本、低信噪比条件下的三维点云目标识别面临挑战。对此提出一种基于形状和骨架特征匹配的目标识别算法，利用语义规则滤波和二维映射，解决杂波干扰、目标点云模糊的问题。设计一种基于质心的编码方式对目标形状和骨架特征进行统一表征，利用组合判决指标实现目标识别。利用室内场景8类家具三维点云进行目标识别仿真实验，结果表明所提算法识别性能优于骨架匹配和形状上下文匹配算法，具备可行性。

关键词: 特征匹配, 三维点云, 目标识别, 小样本

Abstract:

In non-cooperative or obscured scenes, it is difficult to obtain high-quality three-dimensional （3D） point clouds, and the recognition of 3D point clouds under the condition of small samples and low signal-to-noise ratio faces challenges. Considering this, a target recognition algorithm based on shape and skeleton feature matching is proposed to solve the problems of clutter interference and target point cloud ambiguity by using semantic rule filtering and 2D mapping. A coding method based on centroid is designed to represent the shape and skeleton of the target uniformly, and the target recognition is realized by combining decision indexes. An object recognition simulation experiment is carried out by using the 3D point cloud of eight types of furniture in indoor scenes. The results show that the recognition performance of the proposed algorithm is better than that of skeleton matching algorithms and shape context matching algorithms, and it is feasible.

Key words: feature matching, three-dimensional point clouds, target recognition, small sample

中图分类号:

范睿嘉, 刘杰, 于君明, 冯晓峰, 徐文静, 尹良. 基于形状和骨架特征匹配的小样本三维点云目标识别方法[J]. 系统工程与电子技术, 2026, 48(1): 76-86.

Ruijia FAN, Jie LIU, Junming YU, Xiaofeng FENG, Wenjing XU, Liang YIN. Small sample 3D point cloud target recognition method based on shape and skeleton feature matching[J]. Systems Engineering and Electronics, 2026, 48(1): 76-86.

图/表 25

图1

图2

图3

图4

表1

图5

图6

图7

图8

图9

图10

图11

图12

图13

表2

图14

图15

图16

图17

图18

图19

图20

图21

图22

图23

参考文献 25

1	QI S H, NING X, YANG G W, et al. Review of multi-view 3D object recognition methods based on deep learning[J]. Displays, 2021, 69, 102053. doi: 10.1016/j.displa.2021.102053
2	黄奕程. 基于深度学习的三维点云混合数据增强策略研究[D]. 广东: 广东工业大学, 2022.
	HUANG Y C. Research on 3D point cloud mix-up data enhancement strategy based on deep learning[D]. Guangdong: Guangdong University of Technology, 2022.
3	马天恩, 刘涛, 杜萍, 等. 一种聚合全局上下文信息的三维点云语义分割方法[J]. 武汉大学学报(信息科学版), 2025, 50 (12): 2548- 2559.
	MA T E, LIU T, DU P, et al. A 3D point cloud semantic segmentation method for aggregating global context information[J]. Geomatics and Information Science of Wuhan University, 2025, 50 (12): 2548- 2559.
4	朱安迪, 达飞鹏, 盖绍彦. 对融合特征敏感的三维点云识别与分割[J]. 西安交通大学学报, 2024, 58 (5): 52- 63.
	ZHU A D, DA F P, GAI S Y. Recognition and segmentation of point clouds sensitive to fusion features[J]. Journal of Xi’an Jiaotong University, 2024, 58 (5): 52- 63.
5	张健. 基于特征描述的三维点云场景中目标识别方法研究[D]. 郑州: 郑州大学, 2022.
	ZHANG J. Research on object recognition in 3D point cloud scene based on feature description[D]. Zhengzhou: Zhengzhou University, 2022.
6	BOULCH A, GUERRY Y, LE S B, et al. SnapNet: 3D point cloud semantic labeling with 2D deep segmentation networks[J]. Computers & Graphics, 2017, 71, 189- 198.
7	PUI C Y, GHIO S, NG B, et al. Robust 3D ISAR ship classification [C] // Proc. of the IEEE Radar Conference, 2023.
8	DEBES C, AMIN M G, ZOUBIR A M. Target detection in single- and multiple-view through-the-wall radar imaging[J]. IEEE Trans. on Geoscience and Remote Sensing, 2009, 47 (5): 1349- 1361. doi: 10.1109/TGRS.2009.2013460
9	徐守坤, 张路军, 石林, 等. 意图注意力引导的小样本3D点云目标检测[J]. 计算机工程, 2024, 50(12): 288−295.
	XU S K, ZHANG L J, SHI L, et al. Few-shot 3D point cloud object detection guided by intention-attention[J]. Computer Engineering, 2024, 50(12): 288−295.
10	HE S H, WANG Z J, WANGN Y G. CAsgraph: a cascade attention graph network using kernel density estimation based on LIDAR point clouds for 3D object detection[C] // Proc. of the 7th Asian Conference on Artificial Intelligence Technology, 2023: 266−272.
11	武斌, 刘溢安, 赵洁. 结合空间结构卷积和注意力机制的三维点云分类网络[J]. 中国图象图形学报, 2024, 29 (2): 520- 532. doi: 10.11834/jig.230137
	WU B, LIU Y A, ZHAO J. Classification network for 3D point cloud based on spatial structure convolution and attention mechanism[J]. Journal of Image and Graphics, 2024, 29 (2): 520- 532. doi: 10.11834/jig.230137
12	LIN Y P, YEH Y M, CHOU Y C, et al. Attention EdgeConv for 3D point cloud classification[C] // Proc. of the Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, 2021: 2018−2022.
13	HAN Z Z, LU H L, LIU Z B, et al. 3D2SeqViews: aggregating sequential views for 3D global feature learning by CNN with hierarchical attention aggregation[J]. IEEE Trans. on Image Processing, 2019, 28 (8): 3986- 3999. doi: 10.1109/TIP.2019.2904460
14	李乔. 基于深度学习的穿墙雷达目标检测算法研究[D]. 西安: 西安电子科技大学, 2023.
	LI Q. Research on target detection algorithm of through-wall radar based on deep learning[D]. Xi’an: Xidian University, 2023.
15	KAMANI M M, FARHAT F, WISTAR S, et al. Skeleton matching with applications in severe weather detection[J]. Applied Soft Computing, 2018, 70, 1154- 1166. doi: 10.1016/j.asoc.2017.05.037
16	NUGROHO G A, DARMAWAN S, AINA B F, et al. Automatic bow echo shape detection using combination of image processing and skeletonization in X-band radar image[C] // Proc. of the International Conference on Radar, Antenna, Microwave, Electronics, and Telecommunications, 2023: 380−385.
17	YIN J J Q, WANG S Y, LI F M. Division-of-focal-plane polarization image denoising algorithm based on improved principal component analysis[J]. Acta Optica Sinica, 2021, 41 (7): 64- 73.
18	肖骏, 梁步阁, 杨德贵, 等. 穿墙雷达墙体参数估计以及补偿方法综述[J]. 无线电工程, 2022, 52 (11): 2023- 2034.
	XIAO J, LIANG B G, YANG D G, et al. Review of wall parameter estimation and compensation methods of TWR[J]. Radio Engineering, 2022, 52 (11): 2023- 2034.
19	ZHANG T Y, SUEN C Y. A fast parallel algorithm for thinning digital patterns[J]. Communications of the ACM, 1984, 27 (3): 236- 239. doi: 10.1145/357994.358023
20	BAI X, LATECKI L J, Discrete skeleton evolution[C]// Proc. of the 6th International Conference on Energy Minimization Methods in Computer Vision and Pattern Recognition, 2007: 362–374.
21	BERG A C. Shape matching and object recognition using shape contexts[J]. IEEE Trans. on Pattern Analysis and Machine Intelligence, 2002, 24 (4): 509- 522. doi: 10.1109/34.993558
22	VISHWANATH K V, GUPTA D, VAHDAT A, et al. ModelNet: towards a datacenter emulation environment [C]// Proc. of the IEEE 9th International Conference on Peer-to-Peer Computing, 2009: 81−82.
23	VISHWAKARMA S, UMMALANENI V, IQBAL M S, et al. Mitigation of through-wall interference in radar images using denoising autoencoders[C] // Proc. of the IEEE Radar Conference, 2018: 1543−1548.
24	ROLDAN I, FIORANELLI F, YAROYOY A. Total variation compressive sensing for 3D shape estimation in short-range imaging radars[J]. IEEE Trans. on Radar Systems, 2023, 1, 583- 592. doi: 10.1109/TRS.2023.3322630
25	BARZEGAR A S, CHELDAVI A, SEDIGHY S H, et al. 3-D through-the-wall radar imaging using compressed sensing[J]. IEEE Geoscience and Remote Sensing Letters, 2022, 19

特征类型	表征能力	抗噪声能力	获取过程
形状	表征目标外在轮廓，对复杂图像细节表征能力强	对噪声和姿态变化敏感性较高	获取困难较小，计算量较小
骨架	表征目标内部拓扑结构，在一定变化目标观测角度下，骨架主干连接关系保持不变	对噪声和姿态变化敏感性较低	获取质量受剪枝算法影响较大，且计算量较大

皮尔逊线性相关系数	肯德尔秩相关系数	斯皮尔曼秩相关系数
0.6112	0.4483	0.6352

[1]	蒋明煜, 张顺生, 肖思瑶. 面向轻量级交叉注意力卷积网络的SAR目标识别[J]. 系统工程与电子技术, 2025, 47(9): 2853-2861.
[2]	陈宇鹏, 黄科举, 刘辉, 邝龙坤, 杨俊安. 基于信号递归图和卷积宽度学习的小样本辐射源个体识别方法[J]. 系统工程与电子技术, 2025, 47(9): 3086-3092.
[3]	张新征, 闫梦可, 朱晓林. 噪声伪标签容忍的半监督SAR目标识别[J]. 系统工程与电子技术, 2025, 47(6): 1796-1805.
[4]	王彩云, 贾一帆, 李晓飞, 王佳宁, 吴钇达. 基于改进鲸鱼算法和BiGRU的弹道目标HRRP识别方法[J]. 系统工程与电子技术, 2025, 47(6): 1824-1832.
[5]	姜智杰, 宋恒, 胡楠, 段兰茜, 曹平. 隧道环境毫米波雷达目标识别与分类算法[J]. 系统工程与电子技术, 2025, 47(5): 1453-1460.
[6]	郭泽坤, 杨洪飞, 刘峥, 谢荣, 冉磊, 李嘉楠. 基于MGMD空间的窄带雷达空中目标识别方法[J]. 系统工程与电子技术, 2025, 47(4): 1136-1145.
[7]	李家宽, 冯博, 刘红亮, 叶春茂, 余继周. 基于角度引导注意力的气动目标宽带PD识别方法[J]. 系统工程与电子技术, 2025, 47(3): 807-816.
[8]	李娇, 彭文胜, 曾照洋, 邵英华, 张宝良. 小偏度数据下通用质量特性评估最小样本量确定方法[J]. 系统工程与电子技术, 2025, 47(10): 3378-3388.
[9]	蔡伟, 王鑫, 蒋昕昊, 杨志勇, 陈栋. 基于解耦的小样本目标检测方法研究[J]. 系统工程与电子技术, 2024, 46(9): 2941-2950.
[10]	尹建国, 盛文, 蒋伟. 基于深度残差收缩网络的雷达空中目标识别[J]. 系统工程与电子技术, 2024, 46(9): 3012-3018.
[11]	王彩云, 张慧雯, 王佳宁, 吴钇达, 常韵. 基于DTCWT-VAE的弹道中段目标RCS识别[J]. 系统工程与电子技术, 2024, 46(7): 2269-2275.
[12]	吴钇达, 王彩云, 王佳宁, 李晓飞. 基于ISVM-DS的红外多传感器融合识别方法[J]. 系统工程与电子技术, 2024, 46(5): 1555-1560.
[13]	欧阳彤, 汪玲, 朱岱寅, 李勇. 融合LightGBM的ResNeXt气象目标细粒度识别方法[J]. 系统工程与电子技术, 2024, 46(12): 4034-4043.
[14]	王彩云, 常韵, 李晓飞, 王佳宁, 吴钇达, 张慧雯. 基于改进MKELM的红外空间锥体目标识别[J]. 系统工程与电子技术, 2024, 46(10): 3257-3264.
[15]	王湖升, 陈伯孝, 叶倾知. 基于箔条干扰实测数据的对抗方法研究[J]. 系统工程与电子技术, 2023, 45(7): 2010-2021.