博狗体育网站

當前位置: 博狗 >> 科學研究 >> 學術講座 >> 正文
【海韻講座】2019年第32期-基於激光雷達的三維目標檢測及其在無人駕駛中的應用
發佈時間:2019年05月15日 瀏覽次數:

講座人:李軍,博狗体育网站/滑鐵盧大學教授

時間:2019年5月16日,星期四,19:00

地點:海韻行政樓C505

講座摘要:無人駕駛是當前計算機視覺與信號與信息處理學科研究領域最具吸引力和最具挑戰性的研究課題。未來,無論在L4和L5級無人駕駛車輛亟需的環境感知與高精地圖技術方面,激光雷達均已成爲必不可少車載傳感器的輔助導航產品。然而激光雷達點雲數據的複雜特性,又給環境感知與高精地圖技術的研發帶來極大的挑戰。本講座對這一挑戰進行了系統的分析,根據其團隊近5年的研究成果,較爲全面地介紹了爲解決三維目標檢測這一前沿關鍵科學問題研發的激光雷達點雲數據處理算法,特別是融入AI和深度學習算法的新檢測技術,主要對室外路邊三維柱狀物、車輛、和沒有衛星導航定位信號的地下停車場的三維目標檢測提出了出了新穎可行的解決方案。通過展示在自採和公開點雲數據集上比較多種算法的試驗結果,定量地驗證了自主研發技術的先進性。本講座對從事三維目標檢測和激光雷達點雲數據處理以及關注AI和深度學習算法應用研究的師生具有參考意義。

主講人簡介:Dr. Jonathan Li is a full professor in geomatics and systems design engineering at the University of Waterloo, Canada. He has been the Dean, School of Information Science and Engineering at Xiamen University since 2011 while he takes a partial leave from the University of Waterloo. He is the founding member of Waterloo Artificial Intelligence Institute and the co-founder of two startup companies in Waterloo: Ecopia Tech and WatXtract.ai.

Prof. Li has co-authored more than 400 publications, including those in peer-reviewed IEEE journals and top AI conferences (CVPR, AAAI, IJCAI). He has made significant contributions in the fields of mobile mapping, LiDAR point cloud processing, and 3D object detection by co-editing an ISPRS book “Advances in Mobile Mapping Technology” (2007), co-authoring nearly 70 papers in mobile laser scanning published in ISPRS Journal of Photogrammetry and Remote Sensing, IEEE-TGRS, IEEE-TITS, IEEE-JSTARS, IEEE-GRSL, and PERS. He has received several prestigious awards from ASPRS (USA), RSPSoc (UK), CIG and CRSS (Canada), and Joint ISPRS/IAG/FIG Committee.

Prof. Li serves as Chair of ISPRS WG I/2 on LiDAR, Air- and Spaceborne Optical Sensing (2016-2020) and ICWG I/V on Mobile Scanning and Imaging Systems for two terms (2008-2012; 2012-2016). He also chairs the ICA Commission on Sensor-driven Mapping (2015-2019). He is Associate Editor of IEEE-TITS, IEEE-JSTARS, Canadian Journal of Remote Sensing, and Associate Editor-in-Chief of Sensors.

Top