PLP-SLAM:基于点、线、面特征融合的视觉SLAM方法

发布时间：2018-05-28 22:40

  本文选题：同时定位与地图构建 + 点线面特征融合　； 参考：《机器人》2017年02期

【摘要】：基于点特征的视觉SLAM(同时定位与地图构建)算法存在计算量大、环境存储空间负荷高、定位误差较大的问题,为此,提出了一种基于点、线段、平面特征融合的视觉SLAM算法——PLP-SLAM.在扩展卡尔曼滤波(EKF)框架下,首先利用点特征估计机器人当前位姿,然后构建了基于点、线、平面特征的观测模型,最后建立了带平面约束的线段特征数据关联方法及系统状态更新模型,并利用线段和平面特征描述环境信息.在公开数据集上进行了实验,结果表明,本文PLP-SLAM算法能够成功完成SLAM任务,平均定位误差为2.3 m,优于基于点特征的SLAM方法,并通过基于不同特征的SLAM实验表明了本文提出的点、线、面特征融合的优越性.
[Abstract]:The visual slam (simultaneous location and map construction) algorithm based on point feature has the problems of large computation, high storage space load and large positioning error. Therefore, a new algorithm based on point and line segment is proposed. Visual SLAM algorithm for plane feature Fusion PLP-SLAM. In the framework of extended Kalman filter (EKF), the current position and attitude of the robot are estimated by using the point feature, and then the observation model based on the point, line and plane features is constructed. Finally, a line segment feature data association method with plane constraints and a system state update model are established, and line segments and plane features are used to describe the environmental information. Experiments on the open dataset show that the PLP-SLAM algorithm can successfully accomplish the SLAM task, and the average localization error is 2.3 m, which is better than the SLAM method based on the point feature. The point proposed in this paper is shown by the SLAM experiment based on different features. The superiority of line and surface feature fusion.
【作者单位】： 中国民航大学计算机科学与技术学院;福建省信息处理与智能控制重点实验室(闽江学院);
【基金】：国家自然科学基金(61305107,U1333109) 天津市应用基础与前沿技术研究计划重点项目(14JCZDJC32500) 中央高校基本科研业务费(3122016B006) 福建省信息处理与智能控制重点实验室开放课题(MJUKF201732) 福建省科技厅引导性课题(2015H0031)
【分类号】：TP242;TP391.41
，

本文编号：1948460