计算机科学

计算机科学 ›› 2026, Vol. 53 ›› Issue (9): 209-218.doi: 10.11896/jsjkx.260500015

• 计算机图形学&多媒体 • 上一篇    下一篇

融合频域与几何特征的六自由度位姿估计方法

何钰煌1,3, 王乾垒1,3, 戴允童1,3, 杜戊2,4, 秦小林1,3   

  1. 1 中国科学院成都计算机应用研究所 成都 610213
    2 四川大学计算机学院 成都 610065
    3 中国科学院大学计算机科学与技术学院 北京 100049
    4 成都三是科技有限公司 成都 610052
  • 收稿日期:2026-05-06 修回日期:2026-07-05 出版日期:2026-09-15 发布日期:2026-09-10
  • 通讯作者: 秦小林(qinxl2001@126.com)
  • 作者简介:(heyuhuang24@mails.ucas.ac.cn)
  • 基金资助:
    四川省科技计划(2025JDDQ0008,2024NSFJQ0035)

Six Degrees of Freedom Pose Estimation Method Fusing Spectral and Geometric Features

HE Yuhuang1,3, WANG Qianlei1,3 DAI Yuntong1,3 DU Wu2,4, QIN Xiaolin1,3   

  1. 1 Chengdu Institute of Computer Application,Chinese Academy of Sciences,Chengdu 610213,China
    2 School of Computer Science,Sichuan University,Chengdu 610065,China
    3 School of Computer Science and Technology,University of Chinese Academy of Sciences,Beijing 100049,China
    4 Chengdu Sanshi Technology Company Limited,Chengdu 610052,China
  • Received:2026-05-06 Revised:2026-07-05 Published:2026-09-15 Online:2026-09-10
  • About author:HE Yuhuang,born in 2001,postgra-duate.His main research interests include pose estimation and computer vision.
    QIN Xiaolin,born in 1980,Ph.D,researcher,is a senior member of CCF(No.12344M).His main research interests include automated reasoning and spatial intelligence,industry-specific vertical large models.
  • Supported by:
    Sichuan Science and Technology Program(2025JDDQ0008,2024NSFJQ0035).

PDF (PC)

238

摘要： 针对六自由度位姿估计在复杂场景下存在弱纹理与对称目标带来的特征不足与位姿歧义问题,提出了一种融合频域特征增强与几何增强的位姿估计网络SGF-Pose。该方法首先在特征提取阶段引入频域特征增强模块(Spectral Feature Enhancement Module,SFEM),通过学习频谱掩膜自适应地抑制背景高频噪声,以提升预测的鲁棒性;进而,设计了几何增强模块(Geometry Enhancement Module,GEM),通过计算坐标映射的空间梯度显式衡量像素级几何曲率;最后,通过可微的PnP解算模块,结合稠密预测图和几何置信度预测出6D位姿。实验结果表明,采用ADD-(S)指标,SGF-Pose在LineMOD数据集上达到了96.9%的精度,相较于DPOD算法提升了1.7个百分点;在Occlusion LineMOD数据集上达到了59.5%的精度,相较于GDR-Net提升了3.4个百分点;采用AR指标,SGF-Pose在T-LESS数据集上达到了66.4%的精度,相较于CosyPose算法提升了2.4个百分点。可见,SGF-Pose对弱纹理和对称性场景下的位姿估计具有较好的性能。

关键词: 位姿估计, 频域分析, 特征融合, 弱纹理目标, 深度学习

Abstract: To solve the problems of insufficient feature and pose ambiguity caused by texture-less and symmetric objects in complex scenes during 6D pose estimation,a pose estimation network named SGF-Pose integrating frequency-domain feature enhancement and geometric enhancement is proposed.Firstly,a spectral feature enhancement module(SFEM) is introduced in the feature extraction stage.By learning a spectral mask,background high-frequency noise is adaptively suppressed to improve the robustness of the predictions.Furthermore,a geometry enhancement module(GEM) is designed to explicitly measure pixel-level geometric curvature by calculating the spatial gradients of coordinate mappings.Finally,the 6D pose is predicted through a differentiable PnP solver module by combining dense prediction maps and geometric confidence.Experimental results show that,using the ADD-(S) metric,SGF-Pose achieves an accuracy of 96.9% on the LineMOD dataset,which is 1.7 percentage points higher than that of DPOD.On the Occlusion LineMOD dataset,an accuracy of 59.5% is achieved,representing an improvement of 3.4 percentage points over GDR-Net.Using the AR metric,an accuracy of 66.4% is achieved by SGF-Pose on the T-LESS dataset,which is 2.4 percentage points higher than that of CosyPose.These results indicate that SGF-Pose performs well in pose estimation for weak-texture and symmetric scenes.

Key words: Pose estimation, Frequency domain analysis, Feature fusion, Texture-less object, Deep learning

中图分类号: 

  • TP391.41
[1] 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.
[2] GUO N,LI J Y,REN X.Survey of rigid object pose estimation algorithms based on deep learning[J].Computer Science,2023,50(2):178-189.
[3] DORUK A E,OZKAYA T E,GÜLMEZ F,et al.A comparative study for 6D pose estimation of textureless and symmetric objects used in automotive manufacturing industry[C] //2023 5th International Congress on Human-Computer Interaction,Optimization and Robotic Applications.Piscataway:IEEE,2023:1-7.
[4] WANG Y N,JIANG Y M,JIANG J,et al.Key technologies of robot perception and control and its intelligent manufacturing applications[J].Acta Automatica Sinica,2023,49(3):494-513.
[5] GUAN J,HAO Y M,WU Q X,et al.A survey of 6DoF object pose estimation methods for different application scenarios[J].Sensors,2024,24(4):1076.
[6] DI Y,MANHARDT F,WANG G,et al.SO-Pose:exploitingself-occlusion for direct 6D pose estimation[C] //2021 IEEE/CVF International Conference on Computer Vision.Piscataway:IEEE,2021:12376-12385.
[7] WANG G,MANHARDT F,TOMBARI F,et al.GDR-Net:geometry-guided direct regression network for monocular 6D object pose estimation[C] //2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE,2021:16606-16616.
[8] PENG S D,ZHOU X W,LIU Y,et al.PVNet:pixel-wise voting network for 6DoF object pose estimation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2022,44(6):3212-3223.
[9] LIAN R Y,LING H B.CheckerPose:progressive dense keypoint localization for object pose estimation with graph neural network[C] //2023 IEEE/CVF International Conference on Computer Vision.Piscataway:IEEE,2023:13976-13987.
[10] XU L,QU H X,CAI Y J,et al.6D-Diff:a keypoint diffusion framework for 6D object pose estimation[C] //2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE,2024:9676-9686.
[11] YU L X,CHEN Y B,QU H J,et al.Multi-stage grasping method for unordered mixed objects grasping based on GraspNet[J].Computer Science,2026,53(4):318-325.
[12] XU M Z,SHI X H.6D pose estimation based on multi feature fusion[J].Computer Applications and Software,2025,42(6):193-201.
[13] HODAN T,HALUZA P,OBDRZALEK S,et al.T-LESS:anRGB-D dataset for 6D pose estimation of texture-less objects[J].arXiv:1701.05498,2017.
[14] ZHAO H,WEI S X,SHI D H,et al.Learning symmetry-aware geometry correspondences for 6D object pose estimation[C] //2023 IEEE/CVF International Conference on Computer Vision.Piscataway:IEEE,2023:13999-14008.
[15] BAY H,TUYTELAARS T,VAN G L.SURF:speeded up robust features[C] //Proceedings of the 9th European Conference on Computer Vision.Berlin:Springer,2006:404-417.
[16] CALONDER M,LEPETIT V,STRECHA C,et al.BRIEF:binary robust independent elementary features[C] //Proceedings of the 11th European Conference on Computer Vision.Berlin:Springer,2010:778-792.
[17] RUBLEE E,RABAUD V,KONOLIGE K,et al.ORB:an effi-cient alternative to SIFT or SURF[C] //Proceedings of the 2011 International Conference on Computer Vision.Piscataway:IEEE,2011:2564-2571.
[18] XIANG Y,SCHMIDT T,NARAYANAN V,et al.PoseCNN:a convolution neural network for 6D object pose estimation in cluttered scenes[J].arXiv:1711.00199,2017.
[19] LI Y,WANG G,JI X Y,et al.DeepIM:deep iterative matching for 6D pose estimation[J].International Journal of Computer Vision,2020,128:657-678.
[20] LABBE Y,CARPENTIER J,AUBRY M,et al.CosyPose:consistent multi-view multi-object 6D pose estimation[C] //16th European Conference on Computer Vision.Cham:Springer,2020:574-591.
[21] MOON S P,SON H T,HUR D C,et al.GenFlow:generalizablerecurrent flow for 6D pose refinement of novel objects[C] //2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE,2024:10039-10049.
[22] PERIYASAMY A S,AMINI A,TSATURYAN V,et al.YOLOPose V2:understanding and improving transformer-based 6D pose estimation[J].arXiv:2307.11550,2023.
[23] WANG C,XU D F,ZHU Y K,et al.DenseFusion:6D object pose estimation by iterative dense fusion[C] //2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE,2019:3338-3347.
[24] HE Y S,HUANG H B,FAN H Q,et al.FFB6D:a full flow bidirectional fusion network for 6D pose estimation[C] //2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE,2021:3002-3012.
[25] PARK K,PATTEN T,VINCZE M.Pix2Pose:pixel-wise coordinate regression of objects for 6D pose estimation[C] //2019 IEEE/CVF International Conference on Computer Vision.Piscataway:IEEE,2019:7667-7676.
[26] SU Y Z,SALEH M,FETZER T,et al.ZebraPose:coarse to fine surface encoding for 6DoF object pose estimation[C] //2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE,2022:6728-6738.
[27] HODAN T,BARATH D,MATAS J.EPOS:estimating 6D pose of objects with symmetries[C] //2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE,2020:11700-11709.
[28] ORNEK E P,LABBE Y,TEKIN B,et al.FoundPose:unseenobject pose estimation with foundation features[C] //European Conference on Computer Vision.Cham:Springer,2024:163-182.
[29] LIU L H,LIN J H,LIU Z X,et al.PicoPose:progressive pixel-to-pixel correspondence learning for novel object pose estimation[J].arXiv:2504.02617,2025.
[30] WANG H,SRIDHAR S,HUANG J W,et al.Normalized objectcoordinate space for category-level 6D object pose and size estimation[C] //2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE,2019:2637-2646.
[31] LI Z G,WANG G,JI X Y.CDPN:coordinates-based disentangled pose network for real-time RGB-based 6-DoF object pose estimation[C] //2019 IEEE/CVF International Conference on Computer Vision.Piscataway:IEEE,2019:7677-7686.
[32] TEKIN B,SINHA S N,FUA P.Real-time seamless single shot 6D object pose prediction[C] //2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE,2018:292-301.
[33] ZAKHAROV S,SHUGUROV I,ILIC S.DPOD:6D pose object detector and refiner[C] //2019 IEEE/CVF International Confe-rence on Computer Vision.Piscataway:IEEE,2019:1941-1950.
[34] TAN T,DONG Q L.SMOC-Net:leveraging camera pose forself-supervised monocular object pose estimation[C] //2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE,2023:21307-21316.
[35] CHEN H Z,MANHARDT F,NAVAB N,et al.TexPose:neural texture learning for self-supervised 6D object pose estimation[C] //2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE,2023:4841-4852.
[36] WANG G,MANHARDT F,SHAO J Z,et al.Self6D:self-su-pervised monocular 6D object pose estimation[C] //European Conference on Computer Vision.Cham:Springer,2020:108-125.
[37] IWASE S,LIU X Y,KHIRODKAR R,et al.RePOSE:fast 6D object pose refinement via deep texture rendering[C] //2021 IEEE/CVF International Conference on Computer Vision.Piscataway:IEEE,2021:3283-3292.
[38] CHEN B,CHIN T J,KLIMAVICIUS M.Occlusion-robust object pose estimation with holistic representation[C] //2022 IEEE/CVF Winter Conference on Applications of Computer Vision.Piscataway:IEEE,2022:2223-2233.
[39] SHUGUROV I,ZAKHAROV S,ILIC S.DPODv2:dense correspondence-based 6 DoF pose estimation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2022,44(11):7417-7435.
[40] LABBE Y,MANUELLI L,MOUSAVIAN A,et al.MegaPose:6D pose estimation of novel objects via render & compare[C] //Proceedings of the 6th Conference on Robot Learning.Cambridge,MA:JMLR,2023:715-725.
[41] NGUYEN V N,GROUEIX T,SALZMANN M,et al.Giga-Pose:fast and robust novel object pose estimation via one correspondence[C] //2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE,2024:9903-9913.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市两江新区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发