视觉SLAM初始化IMU算法研究
首发时间:2017-04-13
摘要:近些年视觉惯性SLAM算法研究取得重大成果,能够较精确地计算传感器的增量运动。然而这些方法都缺少闭环,即使重复访问同一个地方,而估计的路径误差还在不断的累积。本文实现了一个视觉SLAM和IMU紧耦合的算法,能够闭环和在已成图的区域定位零飘逸。虽然对于单目相机都存在尺度不确定问题,但是可以利用单目SLAM来初始化IMU,能够计算尺度、重力方向、速度、加速度和陀螺仪的偏置。从而后续就可以把视觉SLAM和IMU紧耦合,用来传感器的定位和环境建图。
For information in English, please click here
Initializing IMU with Visual SLAM
Abstract:In recent years there have been excellent results in Visual-Inertial Odometry techniques, which aim to compute the incremental motion of the sensor with high accuracy and robustness. However these approaches lack the capability to close loop, and trajectory estimation accumulates drift even if the sensor is continually revisiting the same place. In this work we present a noveltightly-coupled Visual-Inertial Simultaneous Localization and Mapping system that is able to close loops and reuse its map to achieve zero-drift localization in already mapped areas. We address here that the most general problem of monocular camera, with its well-knowscale ambiguity. We also propose a novel a novel IMU initialization method, which compute the scale, the gravity direction, the velocity,and gyroscope and accelerometer biases, in a few seconds with high accuracy. So Visual SLAM and IMU tight-coupled can be used trackingand mapping.
Keywords: Computer Science,SLAM,IMU,Initialization, tight-coupled
基金:
论文图表:
引用
No.4724437119211214****
同行评议
共计0人参与
勘误表
视觉SLAM初始化IMU算法研究
评论
全部评论0/1000