基于模型选择自动初始化的视觉惯导SLAM系统
首发时间:2020-05-19
摘要:单目视觉惯导的SLAM初始化阶段是整个系统重要且脆弱的一部分,需要一个鲁棒的初始化过程以确保系统的可用性。针对初始化过程中容易出现位姿估计不准确而导致整个系统定位漂移的问题,本文在VINS-Mono的系统框架下,提出了一种基于模型选择的自动初始化方法,使SLAM系统可以根据不同场景自动选择合适的模型来完成联合初始化,得到系统初始值。在Euroc公开数据集的实验验证了该方法的有效性,实验结果表明改进后的VINS-Mono系统的精度和鲁棒性都得到了提高。
For information in English, please click here
Automatic Initialization with Visual-Inertial SLAM System Based on Model Selection
Abstract:The initialization stage of monocular visual-inertial SLAM is an important and fragile part of the entire system, which requires a robust initialization process to ensure the availability of the system. Aiming at the problem of inaccurate pose estimation during the initialization process that leads to the drift of the entire system, this paper proposes an automatic initialization method based on model selection under the VINS-Mono system framework, so that the SLAM system can automatically select the appropriate model to complete the joint initialization and get the initial value of the system. Experiments on the Euroc public dataset verify the effectiveness of the method. The experimental results show that the improved VINS-Mono system has improved accuracy and robustness.
Keywords: Computer science SLAM Monocular visual-inertial Automatic initialization
基金:
引用
No.****
动态公开评议
共计0人参与
勘误表
基于模型选择自动初始化的视觉惯导SLAM系统
评论
全部评论0/1000