報告人:夏彥(德國慕尼黑工業大學 博士后研究員)
時間:2024.1.19 15:00-16:30
地點:測繪館206報告廳
報告人簡介
夏彥博士是德國慕尼黑工業大學Computer Vision Group(CVG)博士后研究員,慕尼黑機器學習中心成員,牛津大學Visual Geometry Group(VGG)訪問學者。博士畢業于慕尼黑工業大學,博士期間在牛津大學VGG組聯合培養。現博士后合作導師為機器人領域專家Daniel Cremers教授(全球top10科學家,德國科研最高獎萊布尼茨獎獲得者)。主要研究方向是三維點云處理,機器人和無人駕駛等,在中科院一區/CCF A類(CVPR, ICCV, ISPRS Journal)等刊物上發表學術論文十幾余篇。現擔任歐洲人工智能科學研究項目評審會成員,國際攝影測量與遙感協會WG I/8工作組秘書、中國留德計算機學會自動駕駛委員會專任委員。
報告簡介
As a cornerstone of achieving autonomous driving, 3D understanding of urban street environments has drawn attention increasingly in the fields of photogrammetry, computer vision, and robotics. Point clouds, acquired via mobile laser scanning, can provide detailed 3D information during driving for understanding the 3D urban street environments. In this talk, point cloud based place recognition, 3D vehicle detection and tracking, and 3D shape completion methods are developed to achieve localization of the current scene within an existing 3D map and status analysis of the dynamic vehicles during the driving.