講座題目：Advanced 3D perception for autonomous mobile systems自主移動機器人的先進三維感知

講座時間：2016年3月29日（周二）13：30

講座地點：科研樓322室

主講人：Laurent Kneip 博士

Laurent Kneip is a DECRA-fellow, senior researcher, and lecturer within the Research School of Engineering at the Australian National University, where he collaborates with Prof Richard Hartley and Prof Hongdong Li. He owns a PhD degree from ETH Zurich, where he worked under the supervision of Prof Roland Siegwart and Prof Marc Pollefeys. He is an expert in geometric computer vision, and his work aims at enabling autonomous systems and mobile applications to use cameras for real-time 3D perception of the environment. His research interests include visual odometry, Simultaneous Localization and Mapping (SLAM) and structure-from-motion with single and multi-camera systems, as well as the efficient solution of the more fundamental, underlying algebraic geometry problems. He published more than 30 papers at major international conferences in computer vision and robotics, and he is the main author of the open-source computer vision libraries OpenGV and polyjam.

Laurent Kneip博士是澳大利亞研究委員會（ARC）卓越青年研究獎（DECRA）獲得者，高級研究員，在澳大利亞國立大學工程研究院與著名學者Richard Hartley教授和李宏東教授一起工作。在蘇黎世聯(lián)邦理工大學獲博士學位，師從著名學者Marc Pollefeys教授和Roland Siegwart教授，是幾何計算機視覺專家，他的研究目標是通過攝像機讓自主移動機器人實時感知周圍3D環(huán)境，研究方向包括視覺計程、同步定位與地圖構(gòu)建、基于多攝像機的從運動恢復結(jié)構(gòu)、以及一些代數(shù)幾何基礎問題的有效求解方法。他在計算機視覺和機器人領(lǐng)域的主要國際會議和期刊上發(fā)表論文30余篇，是開源軟件庫OpenGV 和polyjam的主要作者。

內(nèi)容摘要：

The field of computer vision has been making enormous progress these recent years. With many of the fundamental problems considered solved, we now need to start thinking about where to go next, and how all these capabilities can be re-applied to interesting problems in and around autonomous mobile systems. In this talk, I will present some of my most recent contributions to the field of visual Simultaneous Localization and Mapping (SLAM), as well as an outlook onto what I believe to be interesting future research opportunities towards the next-level 3D perception layer for intelligent systems.

?

近年來計算機視覺領(lǐng)域已取得很大進展，很多基礎問題已得到解決，現(xiàn)在需要開始考慮下一步的發(fā)展，以及如何將理論應用到移動機器人系統(tǒng)中。這次講座將介紹在基于視覺的同步定位與地圖構(gòu)建（Visual SLAM）方面的一些新近研究成果，以及智能機器人系統(tǒng)下一3D感知層中的一些待研究的有趣問題。