文献类型：期刊文献

中文题名：基于双目视觉SLAM三维信息的建立与问题研究

英文题名：MonocularBased on the Analysis and Realization of Binocular Vision Mapping Process

作者：张宇廷[1];朱敏玲[1];郑治伟[2]

机构：[1]北京信息科技大学计算机学院,北京100101;[2]中国劳动关系学院酒店管理学院,北京100048

年份：2019

卷号：0

期号：S01

起止页码：314

中文期刊名：中国软科学

外文期刊名：China Soft Science

收录：CSTPCD、、CSSCI2019_2020、BDHX2017、CSCD2019_2020、BDHX、CSSCI、CSCD

基金：北京市大学生科技创新项目资助(5101923400);促进高校内涵发展-科研水平提高-2019年重点研究培育延续支持项目(NO.5211910957).

语种：中文

中文关键词：ORB-SLAM;双目视觉;点云图;深度图

外文关键词：ORB-SLAM;binocular vision;Point cloud;dePth maP

中文摘要：由于数字图像处理和计算机视觉技术的迅速发展,同步定位与地图创建(Simultaneous Localization and Mapping,SLAM)受到越来越广泛地关注.研究者为避免超声或红外传感器感知信息量有限,鲁棒性差的缺点,采用摄像机作为全自主用移动机器人的感知传感器,使用双目视觉SLAM技术进行即时定位与建图.例如在发生重大灾害的时候,所有外界通讯中断,这时可以使用带摄像头的无人机,在没有环境先验信息的情况下,通过自身搭载的传感器,于运动过程中建立所处环境的地图模型,达到探索地貌,了解灾情,及时抢救的目的.本文介绍了双目相机成像原理以及使用一种快速特征点提取描述算法ORB-SLAM(Oriented FAST and Rotated BRIEF-simultaneous localization and mapping)来处理双目相机采集的信息,并建立点云地图的方法.最后提出了理论与实际相结合的应用场景,为后续双目视觉领域的研究发展方向提供对策及建议.

外文摘要：Due to the raPid develoPment of digital image Processing and comPuter vision technology,creating a simultaneous localization and maP(SLAM)is getting more and more attention.In order to solve the shortcomings of limited information and Poor robustness of ultrasonic or infrared sensors,researchers have adoPted cameras as the sensing sensors of fully autonomous mobile robots,and have used binocular vision SLAM technique for real-time Positioning and maPPing.For examPle,in the event of a major disaster,all the communication interruPt,then you can use the camera of mobile robot,in the absence of a Priori knowledge about the environment,through its own carrying sensors,was established in the Process of the movement environment maP model,to exPlore landscaPe,understand the situation,the PurPose of timely rescue.This PaPer introduces the PrinciPle of binocular camera imaging and using a FAST feature Point extraction algorithm ORB-SLAM(Oriented FAST and Rotated BRIEF-simultaneous localization and maPPing)to handle the binocular camera to collect information,and to establish the method of Point cloud maP.Finally,the aPPlication scenarios combining theory and Practice are ProPosed to Provide countermeasures and Suggestions for the future research and develoPment of binocular vision.