Live Demonstration: Real-Time Vi-SLAM With High-Resolution Event Camera

Gongyu Yang, Qilin Ye, Wanjun He, Lifeng Zhou, Xinyu Chen, Lei Yu, Wen Yang, Shoushun Chen, Wei Li; Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2019, pp. 0-0

Event camera is bio-inspired vision sensors that output pixel-level brightness changes asynchronously. Compared to the conventional frame-based camera, it is with high dynamic range, low latency and high sensitivity, and thus can be exploited in SLAM to tackle the problem of occasions with high-speed camera moving and low-light scenes. In this demo, we implement the visual-inerial SLAM in real time with the recently released event camera, namely, CeleX-V. With the feature of high spatial resolution (1280x800) and low latency <0.5us, the proposed method can provide frames with abundant textures and high time response, which leads to more stable tracking ability and better performance in SLAM system.