现代信息科技2025，Vol.9Issue(20)：35-39,5.DOI:10.19850/j.cnki.2096-4706.2025.20.007
视觉SLAM深度相机无重叠视域下自标定方法研究
Research on Self-calibration Method of Visual SLAM Depth Camera without Overlapping Field of View
摘要
Abstract
Visual sensors are widely used in SLAM(Simultaneous Localization and Mapping)technology.Multi-sensor fusion has become an important research direction in robotic SLAM.Accurate data of visual extrinsic parameters directly affects the success rate of SLAM.However,in practice,extrinsic parameter calibration often contains errors,and sensors and equipment may shift or deform during long-term use,leading to changes in the extrinsic parameters.Meanwhile,calibration methods using calibration plates become difficult when non-overlapping fields of view exist between multiple cameras.This paper studies the calibration of dual-Kinect depth cameras,and utilizes point feature matching relationships to calculate the relative transformation between the two sensors,thereby obtaining the extrinsic parameter calibration results for both cameras.Experimental results show that in environments rich in texture information,the proposed camera self-calibration method yields results consistent with Zhang's calibration method.Moreover,it enables autonomous calibration,offers convenient application,and can meet the demands of practical use.关键词
SLAM/标定/深度相机/相机外参Key words
SLAM/calibration/depth camera/camera extrinsic parameter分类
信息技术与安全科学引用本文复制引用
姜鹏,袁亮,戴立根..视觉SLAM深度相机无重叠视域下自标定方法研究[J].现代信息科技,2025,9(20):35-39,5.基金项目
湖南省教育厅科学研究项目青年专项课题(22B0991) （22B0991）
复杂环境特种机器人控制技术与装备湖南省工程研究中心专项课题(2022HNVITZK003) （2022HNVITZK003）
