Global feature pose estimation method based on keypoint distance

doi:10.11772/j.issn.1001-9081.2025010071

Journal of Computer Applications ›› 2026, Vol. 46 ›› Issue (1): 260-269.DOI: 10.11772/j.issn.1001-9081.2025010071

• Multimedia computing and computer simulation • Previous Articles Next Articles

Global feature pose estimation method based on keypoint distance

Yi XIONG¹, Caiqi WANG¹, Ling MEI², Shiqian WU²()

^1.School of Artificial Intelligence and Automation，Wuhan University of Science and Technology，Wuhan Hubei 430081，China
^2.School of Electronic Information，Wuhan University of Science and Technology，Wuhan Hubei 430081，China

Received:2025-01-20 Revised:2025-04-06 Accepted:2025-04-08 Online:2026-01-10 Published:2026-01-10
Contact: Shiqian WU
About author:XIONG Yi， born in 2000， M. S. candidate. His research interests include intelligent robot.
WANG Caiqi， born in 2000， M. S. candidate. His research interests include intelligent robot.
MEI Ling， born in 1989， Ph. D.， lecturer. His research interests include computer vision， artificial intelligence， machine learning.
Supported by:
National Natural Science Foundation of China(62306218);Natural Science Foundation of Hubei Province(2023AFB070)

基于关键点距离的全局特征位姿估计方法

熊毅¹, 王蔡琪¹, 梅岭², 伍世虔²()

^1.武汉科技大学人工智能与自动化学院，武汉 430081
^2.武汉科技大学电子信息学院，武汉 430081

通讯作者: 伍世虔
作者简介:熊毅（2000—），男，湖北荆州人，硕士研究生，主要研究方向：智能机器人
王蔡琪（2000—），男，湖北荆门人，硕士研究生，主要研究方向：智能机器人
梅岭（1989—），男，湖北武汉人，讲师，博士，主要研究方向：计算机视觉、人工智能、机器学习
基金资助:
国家自然科学基金资助项目(62306218);湖北省自然科学基金资助项目(2023AFB070)

Abstract

Abstract:

To address the problem of low accuracy of pose estimation due to the existence of numerous similar features and non-corresponding points in the point cloud， a global feature pose estimation method based on keypoint distance was proposed. In this method， the global features were constructed using the distances between keypoints， thereby avoiding the influence of similar local features on the accuracy of pose estimation. Meanwhile， to improve the matching speed of global features， a feature matching strategy based on distance comparison table was proposed， so that similarity measurement was carried out on global feature votes through the comparison table， thereby avoiding the interference of non-corresponding points and enhancing the efficiency of finding the correspondences by global features effectively. Finally， these correspondences were subjected to Graph-based Reliability for Outlier Removal （GROR） to eliminate outliers and obtain the transformation pose. Experimental results on four public datasets show that compared with Fast Point Feature Histogram （FPFH）， Signature of Histograms of Orientations （SHOT）， and Binarized Signature of Histograms of Orientations （BSHOT）， the proposed method has the area under the precision-recall curve of the feature matching increased by 116%， 169%， and 137% in average， respectively. Moreover， compared with the original GROR， the proposed method has the rotation error and translation error reduced by 47.38% and 52.43%， respectively.

Key words: machine vision, six degrees of freedom pose estimation, keypoint distance, global feature, distance comparison table

摘要：

为了解决位姿估计中点云存在较多的相似特征和非对应点导致位姿估计精度低的问题，提出一种基于关键点距离的全局特征位姿估计方法。该方法使用关键点之间的距离构建全局特征，避免相似的局部特征对位姿估计精度的影响；同时，为了提升全局特征匹配速度，提出一种基于距离对照表的特征匹配策略，通过对照表对全局特征投票进行相似度量，从而在避免非对应点干扰的同时，有效地提高通过全局特征找寻对应关系的效率。最后，将这些对应关系使用基于对应图可靠性的外点去除策略（GROR）去除外点并得到转换位姿。在4个公开数据集上的实验结果显示，相较于快速点特征直方图（FPFH）、方向直方图签名（SHOT）和二值化方向直方图签名（BSHOT）3个特征描述子，所提方法在特征匹配的精度-召回率曲线下区域面积指标分别平均提升了116%、169%和137%；相较于原GROR，所提方法在旋转误差和平移误差上分别降低了47.38%和52.43%。

关键词: 机器视觉, 六自由度位姿估计, 关键点距离, 全局特征, 距离对照表

CLC Number:

TP391.4

Yi XIONG, Caiqi WANG, Ling MEI, Shiqian WU. Global feature pose estimation method based on keypoint distance[J]. Journal of Computer Applications, 2026, 46(1): 260-269.

熊毅, 王蔡琪, 梅岭, 伍世虔. 基于关键点距离的全局特征位姿估计方法[J]. 《计算机应用》唯一官方网站, 2026, 46(1): 260-269.

Figures/Tables 12

References 32

[1]	GUO J， XING X， QUAN W， et al. Efficient center voting for object detection and 6D pose estimation in 3D point cloud ［J］. IEEE Transactions on Image Processing， 2021， 30： 5072-5084.
[2]	MEI L， LAI J， FENG Z， et al. From pedestrian to group retrieval via Siamese network and correlation ［J］. Neurocomputing， 2020， 412： 447-460.
[3]	MEI L， FU M， WANG B， et al. LSN-GTDA： learning symmetrical network via global thermal diffusion analysis for pedestrian trajectory prediction in unmanned aerial vehicle scenarios ［J］. Remote Sensing， 2025， 17（1）： No.154.
[4]	邴雅星，王阳萍，雍玖，等.基于筛选学习网络的六自由度目标位姿估计算法［J］.计算机应用， 2024， 44（6）： 1920-1926.
	BING Y X， WANG Y P， YONG J， et al. Six degrees of freedom object pose estimation algorithm based on filter learning network ［J］. Journal of Computer Applications， 2024， 44（6）： 1920-1926.
[5]	王一，谢杰，程佳，等.基于深度学习的RGB图像目标位姿估计综述［J］.计算机应用， 2023， 43（8）： 2546-2555.
	WANG Y， XIE J， CHENG J， et al. Review of object pose estimation in RGB images based on deep learning ［J］. Journal of Computer Applications， 2023， 43（8）： 2546-2555.
[6]	HINTERSTOISSER S， LEPETIT V， ILIC S， et al. Model based training， detection and pose estimation of texture-less 3D objects in heavily cluttered scenes［M］. Berlin： Springer， 2012.
[7]	WEN B， YANG W， KAUTZ J， et al. FoundationPose： unified 6D pose estimation and tracking of novel objects ［C］// Proceedings of the 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2024： 17868-17879.
[8]	JIANG H， SALZMANN M， DANG Z， et al. SE （3） diffusion model-based point cloud registration for robust 6D object pose estimation ［C］// Proceedings of the 37th International Conference on Neural Information Processing Systems. Red Hook： Curran Associates Inc.， 2023： 21285-21297.
[9]	CHEN Y， LI Z， LI Q， et al. Pose estimation algorithm based on point pair features using PointNet++ ［J］. Complex and Intelligent Systems， 2024， 10（5）： 6581-6595.
[10]	LIU Q， DING K， ZHANG C， et al. A point cloud matching algorithm based on multiscale point pair features ［C］// Proceedings of the 2023 IEEE International Conference on Real-time Computing and Robotics. Piscataway： IEEE， 2023： 953-958.
[11]	YU S， ZHAI D H， ZHAN Y， et al. 6-D object pose estimation based on point pair matching for robotic grasp detection ［J］. IEEE Transactions on Neural Networks and Learning Systems， 2025， 36（7）： 11902-11916.
[12]	DROST B， ULRICH M， NAVAB N， et al. Model globally， match locally： efficient and robust 3D object recognition ［C］// Proceedings of the 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2010： 998-1005.
[13]	JOHNSON A E， HEBERT M. Using spin images for efficient object recognition in cluttered 3D scenes ［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 1999， 21（5）： 433-449.
[14]	TOMBARI F， SALTI S， DI STEFANO L. Unique shape context for 3D data description ［C］// Proceedings of the 2010 ACM Workshop on 3D Object Retrieval. New York： ACM， 2010： 57-62.
[15]	TOMBARI F， SALTI S， DI STEFANO L. Unique signatures of histograms for local surface description ［C］// Proceedings of the 2010 European Conference on Computer Vision， LNCS 6313. Berlin： Springer， 2010： 356-369.
[16]	RUSU R B， BLODOW N， BEETZ M. Fast Point Feature Histograms （FPFH） for 3D registration ［C］// Proceedings of the 2009 IEEE International Conference on Robotics and Automation. Piscataway： IEEE， 2009： 3212-3217.
[17]	RUSU R B， BRADSKI G， THIBAUX R， et al. Fast 3D recognition and pose using the viewpoint feature histogram ［C］// Proceedings of the 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway： IEEE， 2010： 2155-2162.
[18]	ALDOMA A， VINCZE M， BLODOW N， et al. CAD-model recognition and 6DOF pose estimation using 3D cues ［C］// Proceedings of the 2011 IEEE International Conference on Computer Vision Workshops. Piscataway： IEEE， 2011： 585-592.
[19]	郝雯，汪洋，魏海南.基于多特征融合的点云场景语义分割［J］.计算机应用， 2023， 43（10）： 3202-3208.
	HAO W， WANG Y， WEI H N. Semantic segmentation of point cloud scenes based on multi-feature fusion ［J］. Journal of Computer Applications， 2023， 43（10）： 3202-3208.
[20]	朱新成，何坤金，倪娜，等.基于改进迭代最近点算法的接骨板贴合性快捷计算方法［J］.计算机应用， 2021， 41（10）： 3033-3039.
	ZHU X C， HE K J， NI N， et al. Rapid calculation method of orthopedic plate fit based on improved iterative closest point algorithm ［J］. Journal of Computer Applications， 2021， 41（10）： 3033-3039.
[21]	YANG H， SHI J， CARLONE L. TEASER： fast and certifiable point cloud registration ［J］. IEEE Transactions on Robotics， 2021， 37（2）： 314-333.
[22]	YAN L， WEI P， XIE H， et al. A new outlier removal strategy based on reliability of correspondence graph for fast point cloud registration ［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2023， 45（7）： 7986-8002.
[23]	CURLESS B， LEVOY M. A volumetric method for building complex models from range images ［C］// Proceedings of the 23rd Annual Conference on Computer Graphics and Interactive Techniques. New York： ACM， 1996： 303-312.
[24]	MIAN A S， BENNAMOUN M， OWENS R. Three-dimensional model-based object recognition and segmentation in cluttered scenes ［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2006， 28（10）： 1584-1601.
[25]	TEJANI A， TANG D， KOUSKOURIDAS R， et al. Latent-Class Hough Forests for 3D object detection and pose estimation ［C］// Proceedings of the 2014 European Conference on Computer Vision， LNCS 8694. Cham： Springer， 2014： 462-477.
[26]	GUO Y， BENNAMOUN M， SOHEL F， et al. A comprehensive performance evaluation of 3D local feature descriptors ［J］. International Journal of Computer Vision， 2016， 116（1）： 66-89.
[27]	PRAKHYA S M， LIU B， LIN W. B-SHOT： a binary feature descriptor for fast and efficient keypoint matching on 3D point clouds ［C］// Proceedings of the 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway： IEEE， 2015： 1929-1934.
[28]	ZHOU Q Y， PARK J， KOLTUN V. Fast global registration ［C］// Proceedings of the 2016 European Conference on Computer Vision， LNCS 9906. Cham： Springer， 2016： 766-782.
[29]	XING X， LU Z， WANG Y， et al. Efficient single correspondence voting for point cloud registration ［J］. IEEE Transactions on Image Processing， 2024， 33： 2116-2130.
[30]	MEI L， HE Y， FISHANI F J， et al. Learning domain-adaptive landmark detection-based self-supervised video synchronization for remote sensing panorama ［J］. Remote Sensing， 2023， 15（4）： No.953.
[31]	PAN C， FANG H， ZHANG H， et al. Visual attention-guided weighted Naïve Bayes for behavior intention inference ［C］// Proceedings of the 2nd International Conference on Artificial Intelligence， Human-Computer Interaction and Robotics. Piscataway： IEEE， 2023： 569-574.
[32]	MEI L， LAI J， XIE X， et al. Illumination-invariance optical flow estimation using weighted regularization transform ［J］. IEEE Transactions on Circuits and Systems for Video Technology， 2020， 30（2）： 495-508.

数据集	点云密度质量	挑战	模型数	场景数
B3R	高	高斯噪声	6	45
Random Views	高	遮挡、噪声	6	108
UWA	中	遮挡、杂乱	4	50
Kinect	低	遮挡、杂乱	6	16

数据集	点云密度质量	挑战	模型数	场景数
B3R	高	高斯噪声	6	45
Random Views	高	遮挡、噪声	6	108
UWA	中	遮挡、杂乱	4	50
Kinect	低	遮挡、杂乱	6	16

方法	场景1	场景2	场景3	场景4	场景5	场景6	场景7	场景8	场景9	场景10	场景11	场景12
FGR	0.002 7	0.265 2	0.056 1	0.118 6	0.874 4	0.049 8	Failed	0.004 1	0.018 7	Failed	Failed	Failed
Teaser++	0.001 6	0.007 3	0.007 5	0.669 6	0.169 7	0.026 9	Failed	0.010 5	0.020 5	Failed	Failed	Failed
SCVC	0.001 9	0.005 3	0.019 5	0.023 6	0.035 2	0.0041	Failed	0.139 3	0.022 5	Failed	0.087 1	0.067 3
GROR	0.000 8	0.007 5	0.002 3	0.061 3	0.045 8	0.011 9	Failed	0.009 6	0.0135	0.177 6	0.051 4	0.128 8
SCVCG	0.0005	0.0023	0.0019	0.0126	0.0241	0.003 8	0.0863	0.005 1	0.021 7	0.1596	0.0483	0.0425
本文方法	0.000 0	0.000 0	0.001 2	0.003 1	0.011 4	0.005 2	0.077 2	0.0089	0.009 8	0.148 4	0.042 4	0.038 2

方法	场景1	场景2	场景3	场景4	场景5	场景6	场景7	场景8	场景9	场景10	场景11	场景12
FGR	0.002 7	0.265 2	0.056 1	0.118 6	0.874 4	0.049 8	Failed	0.004 1	0.018 7	Failed	Failed	Failed
Teaser++	0.001 6	0.007 3	0.007 5	0.669 6	0.169 7	0.026 9	Failed	0.010 5	0.020 5	Failed	Failed	Failed
SCVC	0.001 9	0.005 3	0.019 5	0.023 6	0.035 2	0.0041	Failed	0.139 3	0.022 5	Failed	0.087 1	0.067 3
GROR	0.000 8	0.007 5	0.002 3	0.061 3	0.045 8	0.011 9	Failed	0.009 6	0.0135	0.177 6	0.051 4	0.128 8
SCVCG	0.0005	0.0023	0.0019	0.0126	0.0241	0.003 8	0.0863	0.005 1	0.021 7	0.1596	0.0483	0.0425
本文方法	0.000 0	0.000 0	0.001 2	0.003 1	0.011 4	0.005 2	0.077 2	0.0089	0.009 8	0.148 4	0.042 4	0.038 2

方法	场景1	场景2	场景3	场景4	场景5	场景6	场景7	场景8	场景9	场景10	场景11	场景12
FGR	0.009 7	0.261 5	0.736 2	1.028 6	8.924 8	0.199 7	Failed	0.2582	1.083 2	Failed	Failed	Failed
Teaser++	0.014 1	0.262 2	0.073 6	2.592 5	1.823 0	0.049 6	Failed	0.543 3	0.923 1	Failed	Failed	Failed
SCVC	0.017 7	0.020 7	0.100 7	0.144 5	0.418 6	0.048 8	Failed	8.600 2	1.169 8	Failed	7.998 4	6.333 7
GROR	0.0039	0.012 8	0.022 2	0.747 4	0.331 6	0.0236	Failed	0.515 4	0.0078	9.110 6	4.896 4	11.648 7
SCVCG	0.002 2	0.0102	0.0145	0.0472	0.1714	0.039 6	5.2397	0.212 8	0.250 0	8.6845	2.7392	5.5275
本文方法	0.005 4	0.007 9	0.011 7	0.010 3	0.099 7	0.008 3	4.575 9	0.442 9	0.006 1	6.050 5	2.102 2	4.249 9

Global feature pose estimation method based on keypoint distance

基于关键点距离的全局特征位姿估计方法

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 12

References 32

Related Articles 15

Recommended Articles

Metrics

方法	场景1	场景2	场景3	场景4	场景5	场景6	场景7	场景8	场景9	场景10	场景11	场景12
FGR	0.513 2	0.249 6	0.324 8	0.235 5	0.279 2	0.213 8	0.227 3	0.2957	0.2611	3.4796	3.2575	3.3384
Teaser++	0.312 4	0.198 4	0.268 3	0.2623	0.2816	0.2573	0.2315	0.263 4	0.247 7	3.293 2	2.934 8	3.034 8
SCVC	0.782 1	0.317 2	0.369 2	0.475 1	0.471 3	0.439 2	0.544 5	0.521 3	0.542 9	5.462 0	5.381 1	5.631 7
GROR	0.4507	0.2372	0.2801	0.335 8	0.363 8	0.352 8	0.322 3	0.393 6	0.357 6	4.187 4	3.578 6	3.834 3
SCVCG	0.829 1	0.364 8	0.408 8	0.502 9	0.532 8	0.500 1	0.648 2	0.612 8	0.589 0	5.723 9	5.617 3	5.837 8
本文方法	0.517 4	0.273 7	0.318 9	0.375 4	0.405 4	0.383 7	0.463 7	0.464 8	0.426 7	4.911 0	4.012 3	4.226 7

[1]	Zhigang XU, Chuang ZHANG. Multi-level color restoration of mural image based on gated positional encoding [J]. Journal of Computer Applications, 2024, 44(9): 2931-2937.
[2]	Yuwei DING, Hongbo SHI, Jie LI, Min LIANG. Image denoising network based on local and global feature decoupling [J]. Journal of Computer Applications, 2024, 44(8): 2571-2579.
[3]	Rong HUANG, Junjie SONG, Shubo ZHOU, Hao LIU. Image aesthetic quality evaluation method based on self-supervised vision Transformer [J]. Journal of Computer Applications, 2024, 44(4): 1269-1276.
[4]	Cheng WANG, Yang WANG, Yingjiao RONG. YOLOv7-MSBP target location algorithm for character recognition of power distribution cabinet [J]. Journal of Computer Applications, 2024, 44(10): 3191-3199.
[5]	Hanyu ZHANG, Zhenbo LI, Weiran LI, Pu YANG. Review of research on aquaculture counting based on machine vision [J]. Journal of Computer Applications, 2023, 43(9): 2970-2982.
[6]	Jiahong SUI, Yingchi MAO, Huimin YU, Zicheng WANG, Ping PING. Global image captioning method based on graph attention network [J]. Journal of Computer Applications, 2023, 43(5): 1409-1415.
[7]	Xinyu CHEN, Mingzhe LIU, Jun REN, Ying TANG. Parameter asynchronous updating algorithm based on multi-column convolutional neural network [J]. Journal of Computer Applications, 2022, 42(2): 395-403.
[8]	LIU Ziyan, ZHU Mingcheng, YUAN Lei, MA Shanshan, CHEN Lingzhouting. Video person re-identification based on non-local attention and multi-feature fusion [J]. Journal of Computer Applications, 2021, 41(2): 530-536.
[9]	ZHANG Yazhou, LU Xianling. Surface defect detection method of light-guide plate based on improved coherence enhancing diffusion and texture energy measure-Gaussian mixture model [J]. Journal of Computer Applications, 2020, 40(5): 1545-1552.
[10]	GAO Ming, REN Dejun, HU Yunqi, FU Lei, QIU Lyu. Ampoule packaging quality inspection algorithm based on machine vision and lightweight neural network [J]. Journal of Computer Applications, 2020, 40(10): 2899-2903.
[11]	YANG Chunde, LIU Jing, QU Zhong. Fast scale adaptive object tracking algorithm with separating window [J]. Journal of Computer Applications, 2019, 39(4): 1145-1149.
[12]	SHI Changyou, WANG Meili, LIU Xinran, HUANG Huili, ZHOU Deqiang, DENG Ganran. Node recognition for different types of sugarcanes based on machine vision [J]. Journal of Computer Applications, 2019, 39(4): 1208-1213.
[13]	YANG Jidong, SUN Zhaoqi, WANG Feilong. Optimal path control of parallel truss manipulator based on visual grasping [J]. Journal of Computer Applications, 2019, 39(3): 913-917.
[14]	LIU Baocheng, PIAO Yan, TANG Yue. Person re-identification in video sequence based on spatial-temporal regularization [J]. Journal of Computer Applications, 2019, 39(11): 3216-3220.
[15]	ZHANG Yunke, LIU Dan. Real-time implementation of improved TINY YOLO vehicle detection algorithm based on Zynq SoC hardware acceleration [J]. Journal of Computer Applications, 2019, 39(1): 192-198.