单目视觉SLAM车载摄像机快速位姿估计及景物重构

doi:10.6040/j.issn.1671-9352.0.2016.385

山东大学学报（理学版） ›› 2016, Vol. 51 ›› Issue (12): 116-124.doi: 10.6040/j.issn.1671-9352.0.2016.385

• • 上一篇

单目视觉SLAM车载摄像机快速位姿估计及景物重构

杨元慧,李国栋^*,吴春富,王小龙

龙岩学院机电工程学院, 福建龙岩 364012

收稿日期:2016-08-09 出版日期:2016-12-20 发布日期:2016-12-20
通讯作者: 李国栋(1981— ),男,博士,讲师,研究方向为视觉伺服、机器视觉、智能空间. E-mail:liguodong.sdu@foxmail.com E-mail:huihui-80@163.com
作者简介:杨元慧(1980— ),女,讲师,研究方向为智能机器人控制、机器视觉;机械设计制造及自动化. E-mail:huihui-80@163.com
基金资助:
国家自然科学基金资助项目(61375084);山东省自然科学基金重点资助项目(ZR2015QZ08);福建省自然科学基金面上资助项目(2015J01268);福建省教育厅科技计划资助项目(JK2014049);福建省科技厅引导性资助项目(2016H0026);福建省教育厅中青年教师教育科研资助项目(JA15499,JA14307);龙岩学院百名青年教师攀登计划资助项目(LQ2013015,LQ2016006);龙岩学院校级产学研资助项目(LC2014003)

Fast pose estimation for on-board camera and scene reconstruction in monocular vision SLAM

YANG Yuan-hui, LI Guo-dong^*, WU Chun-fu, WANG Xiao-long

School of Mechanical and Electrical Engineering, Longyan University, Longyan 364012, Fujian, China

Received:2016-08-09 Online:2016-12-20 Published:2016-12-20

摘要/Abstract

摘要： 针对单目视觉同时定位与地图构建(simultaneous localization and mapping, SLAM)应用,提出一种快速车载摄像机位姿估计及景物结构3D重构算法。在无具体标定物的情况下,利用移动机器人二自由度运动导致摄像机采集视图中对应极点的特殊性质,标定出真实摄像机与虚拟摄像机间的相对姿态,并利用主动视觉方法进一步标定出移动机器人坐标系与虚拟摄像机坐标系间的相对位移;构造无穷单应变换,将真实视图中通过SIFT算法得到的假设欧氏匹配点集转换为对应虚拟摄像机的虚拟假设欧氏匹配点集,并利用基于RANSAC的归一化三点算法快速地对本质矩阵进行估计和分解;利用已观测到的路标三维信息,递归地剔除本质矩阵分解出的平移量的尺度不确定性,并利用线性三角形法重构出景物结构。实验结果表明:该算法在保证运算精度的同时,具有更快的运算速度。

关键词: 归一化三点算法, 随机抽样一致, 对极几何, 本质矩阵分解, 视觉同时定位与地图创建

Abstract: According to the simulaneous localization and mapping(SLAM), a fast pose estimation of on-board camera, together with 3D reconstruction of scene structure algorithm was proposed. The special properties of the Euclidean epipoles corresponding to the mobile robots 2-DOF movement were utilized to calibrate the relative pose information between real camera coordinate and virtual camera coordinate without the utilization of specific calibration object, and the active vision method was utilized to further calibrate the relative position information between mobile robot coordinate and virtual camera coordinate; An constructed infinite homography was adopted to turn the hypothesis Euclidean point correspondences obtained by SIFT algorithm into the virtual hypothesis Euclidean point correspondences, and the RANSAC based normalized 3-point algorithm was implemented to estimate and decompose the essential matrix; The previous 3D information of the observed landmarks were adopted to eliminate the scale uncertainty of the translation vector acquired by essential matrix decomposition, and the scene structure was reconstructed by linear triangulation method. Experimental results show that the proposed algorithm has the advantages of high precision, as well as the low computational complexity.

Key words: essential matrix decomposition, random sample consensus, epipolar geometry, vision SLAM, normalized 3-point algorithm

中图分类号:

TP242

杨元慧,李国栋,吴春富,王小龙. 单目视觉SLAM车载摄像机快速位姿估计及景物重构[J]. 山东大学学报（理学版）, 2016, 51(12): 116-124.

YANG Yuan-hui, LI Guo-dong, WU Chun-fu, WANG Xiao-long. Fast pose estimation for on-board camera and scene reconstruction in monocular vision SLAM[J]. JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE), 2016, 51(12): 116-124.

参考文献

[1] 陈华,王朝立,杨芳, 等. 基于视觉伺服非完整移动机器人的有限时间饱和镇定[J]. 控制理论与应用,2012,29(6):817-823. CHEN Hua, WANG Chaoli, YANG Fang, et al. Finite-time saturated stabilization of nonholonomic mobile robots based on visual servoing[J]. Journal of Control Theory and Application, 2012, 29(6):817-823.
[2] 李宝全,方勇纯,张雪波. 基于2D三焦点张量的移动机器人视觉伺服镇定控制[J]. 自动化学报,2014, 40(12):2706-2715. LI Baoquan, FANG Yongchun, ZHANG Xuebo. 2D Trifocal tensor based visual servo regulation of nonholonomic mobile robots[J]. ACTA Automatica Sinica, 2014, 40(12):2706-2715.
[3] SCHMIDT A, KRAFT M, FULARZ M, et al. On augmenting the visual SLAM with direct orientation measurement using the 5-point algorithm[J]. Journal of Automation Mobile Robotics and Intelligent Systems, 2013, 103(1):5-10.
[4] TRIBOU M J, WASLANDER S L, WANG D W L. Scale recovery in multicamera cluster SLAM with non-overlapping fields of view[J]. Computer Vision and Image Understanding, 2016, 50(C):27-41.
[5] AN S Y, LEE L K, OH S Y. Ceiling vision-based active SLAM framework for dynamic and wide-open environments [J]. Autonomous Robots, 2016, 40(2):291-324.
[6] LONGUET-HIGGINS H C. A computer algorithm for reconstructing a scene from two projections[J]. Nature, 1981, 293:133-135.
[7] HUANG T S, FAUGERAS O D. Some properties of the E matrix in two-view motion estimation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1989, 11(12):1310-1312.
[8] HORN B K P. Relative orientation[J]. International Journal of Computer Vision, 1990, 4(1):59-78.
[9] HARTLEY R I. In defense of the eight-point algorithm[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1997, 19(6):580-593.
[10] DIXON W E, DAWSON D M, ZERGEROGLU E, et al. Adaptive tracking control of a wheeled mobile robot via an uncalibrated camera system[J]. IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics, 2001, 31(3):341-352.
[11] 王伟, 吴成柯. 估计基础矩阵的六点综合算法[J]. 中国科学(E辑), 1997, 27(2):165-170. WANG Wei, WU Chengke. 6-point synthesis algorithm for fundamental matrix estimation[J]. Science of China(E series), 1997, 27(2):165-170.
[12] PHILIP J. A non-iterative algorithm for determining all essential matrices corresponding to five point pairs[J]. The Photogrammetric Record, 1996, 15(88):589-599.
[13] NISTER D. An efficient solution to the five-point relative pose problem[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2004, 26(6):756-770.
[14] STEWENIUS H, ENGELS C, NISTÉR D. Recent developments on direct relative orientation[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2006, 60(4):284-294.
[15] FAUGERAS O D. What can be seen in three dimensions with an uncalibrated stereo rig[C] //Proceedings of the 2nd European Conference on Computer Vision. Genoa, Italy: The Computer Vision Society, 1992:563-578.
[16] MA Y, KOSECKÁ J, SASTRY S. Optimization criteria and geometric algorithms for motion and structure estimation[J]. International Journal of Computer Vision, 2001, 44(3):219-249.
[17] HELMKE U, HÜPER K, LEE P Y, et al. Essential matrix estimation using Gauss-Newton iterations on a manifold[J]. International Journal of Computer Vision, 2007, 74(2): 117-136.
[18] FAUGERAS O.Three dimensional computer vision: a geometric viewpoint[M]. Massachusetts: the MIT Press, 1993.
[19] HARTLEY R, ZISSERMAN A. Multiple view geometry in computer vision [M]. Cambridge: Cambridge University Press, 2000.
[20] MA Y, SOATTO S, KOSECKA J. An invitation to 3-d vision: from images to geometric models[M]. Berlin: Springer, 2004.

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed

单目视觉SLAM车载摄像机快速位姿估计及景物重构

Fast pose estimation for on-board camera and scene reconstruction in monocular vision SLAM

PDF (PC)

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 2

多维度评价

本文评价

推荐阅读 0

[1]	唐庆顺, 吴春富, 李国栋, 王小龙, 周风余. 移动机器人车载摄像机位姿的高精度快速求解[J]. 山东大学学报（理学版）, 2015, 50(03): 32-39.
[2]	吴春富1,唐庆顺1,谢煌生1,周风余2*. 一种新型的本质矩阵解析分解算法[J]. 山东大学学报（理学版）, 2014, 49(03): 31-36.