This website requires JavaScript.

ESVIO: Event-based Stereo Visual Inertial Odometry

Peiyu ChenWeipeng GuanPeng Lu
Dec 2022
摘要
Event cameras that asynchronously output low-latency event streams providegreat opportunities for state estimation under challenging situations. Despiteevent-based visual odometry having been extensively studied in recent years,most of them are based on monocular and few research on stereo event vision. Inthis paper, we present ESVIO, the first event-based stereo visual-inertialodometry, which leverages the complementary advantages of event streams,standard images and inertial measurements. Our proposed pipeline achievestemporal tracking and instantaneous matching between consecutive stereo eventstreams, thereby obtaining robust state estimation. In addition, the motioncompensation method is designed to emphasize the edge of scenes by warping eachevent to reference moments with IMU and ESVIO back-end. We validate that bothESIO (purely event-based) and ESVIO (event with image-aided) have superiorperformance compared with other image-based and event-based baseline methods onpublic and self-collected datasets. Furthermore, we use our pipeline to performonboard quadrotor flights under low-light environments. A real-worldlarge-scale experiment is also conducted to demonstrate long-termeffectiveness. We highlight that this work is a real-time, accurate system thatis aimed at robust state estimation under challenging environments.
展开全部
图表提取

暂无人提供速读十问回答

论文十问由沈向洋博士提出,鼓励大家带着这十个问题去阅读论文,用有用的信息构建认知模型。写出自己的十问回答,还有机会在当前页面展示哦。

Q1论文试图解决什么问题?
Q2这是否是一个新的问题?
Q3这篇文章要验证一个什么科学假设?
0
被引用
笔记
问答