一种单目VIO定位精度与跟踪稳定性优化方法

Optimization method for positioning accuracy and tracking stability of monocular VIO

为了提高视觉惯性同时定位与建图(visual-inertial simultaneous localization and mapping,VISLAM)系统的系统性能,提出了一种单目视觉惯性里程计(visual-inertial odometry,VIO)定位精度与跟踪稳定性优化方法。在相机位姿优化阶段,通过多残差项对相机位姿进行优化,提高系统的定位精度。在特征跟踪丢失时,通过惯性测量单元(inertial measurement unit,IMU)积分信息与特征点深度增强方法对系统进行重定位,提高系统的跟踪稳定性。针对所提方法,基于ORB-SLAM3代码框架进行改进,并在开源数据集与真实环境中验证方法的可行性。实验结果表明,所提方法能够有效提高视觉惯性里程计的定位精度与跟踪稳定性。

In order to improve the system performance of visual-inertial simultaneous localization and mapping(VISLAM)system,an optimization method for positioning accuracy and tracking stability of monocular visual-inertial odometry(VIO)is proposed.In the camera pose optimization stage,in order to improve the positioning accuracy of the system,the camera pose is optimized through multiple residual items.When feature tracking is lost,in order to improve the tracking stability of the system,the integral information of inertial measurement unit(IMU)and the depth enhancement method of feature point are used to relocate the system.The proposed method is improved based on the ORB-SLAM3 code framework,and the feasibility of the method is verified in open source data sets and real environments.The experimental results show that the proposed method can effectively improve the positioning accuracy and tracking stability of the visual-inertial odometry.