摘要
针对计算机视觉中传统ICP算法(Iterative Closest Point)位姿估计的噪音问题,提出一种融合ICP算法和EPnP算法(Efficient Perspective-n-Point)的位姿估计。首先通过分析比较ICP算法、EPnE算法的优缺点,然后针对ICP算法对点云特征点未知情况下的失真问题,将EPnE算法对噪音处理的优势引入ICP算法中,提出了ICP和EPnE融合算法,结果表明上述算法在未知深度下对点云的识别更为准确,验证了上述算法的有效性,为机器视觉技术在人工智能产业的应用提供参考。
The paper proposed a pose estimation approach combining ICP algorithm and EPNP algorithm to ad-dress the noises of the traditional ICP algorithm for pose estimation in computer vision.A comparison between ICP al-gorithm and EPNE algorithm was analysed firstly.Then,the fusion algorithm which introduces the advantages of noise processing from EPnE algorithm into ICP algorithm was presented to deal with the distortion problem of ICP algorithm in the case of unknown point cloud position.The results show that this fusion algorithm is more accurate in identifying the point clouds at unknown position,which verifies the effectiveness of the above algorithm and provides a reference for the application of machine vision technology in the artificial intelligence industry.
作者
孟建军
陈晓彤
李德仓
祁文哲
MENG Jian-jun;CHEN Xiao-tong;LI De-cang;QI Wen-zhe(Mechatronics T&R Institute,Lanzhou Jiaotong University,Lanzhou Gansu 730070,China;Gansu Provincial Engineering Technology Center for Informatization of Logistics&Transport Equipment,Lanzhou Gansu 730070,China;Gansu Provincial Industry Technology Center of Logistics&Transport Equipment,Lanzhou Gansu 730070,China)
出处
《计算机仿真》
北大核心
2023年第5期274-278,共5页
Computer Simulation
基金
国家自然科学基金资助项目(72061021,62063013)
甘肃省高等学校科研项目:(2017D-09,2018C-10)
甘肃省科技计划资助项目(20JR10RA251)。
关键词
计算机视觉
位姿估计
算法
Computer vision
Pose estimation
Algorithm