講演抄録/キーワード |
講演名 |
2018-06-12 13:55
Towards ADAS Map: Automatic Extraction and Integration of Semantic Traffic Regulation from MMS Data ○Yue Zhang・Ehsan Javanmardi・Mahdi Javanmardi・Yanlei Gu・Shunsuke Kamijo(UTokyo) ITS2018-2 |
抄録 |
(和) |
(まだ登録されていません) |
(英) |
Advanced Driver Assistance Systems (ADAS) map is one of the most important components for autonomous driving. And traffic facilities such as road markings, traffic lights and traffic signs which provide semantic information (traffic regulations) for vehicles, are essential parts of the ADAS map. However, traditional methods add the semantic information into map in a manual way, which are costly and time-consuming. Toward ADAS map, we present a framework for extraction and integration of semantic traffic regulation from Mobile Mapping System (MMS) data automatically. In this paper, we focus on the road marking extraction and traffic light detection by means of deep learning and point cloud intensity information from the data recorded by MMS. Experiments have been conducted around Hitotsubashi intersection, a dense urban area of Tokyo, Japan. The results show the effectiveness of our proposed solution and potential capability for mapping. |
キーワード |
(和) |
/ / / / / / / |
(英) |
ADAS map / Point cloud / Deep learning / Semantic traffic regulation / / / / |
文献情報 |
信学技報, vol. 118, no. 79, ITS2018-2, pp. 7-12, 2018年6月. |
資料番号 |
ITS2018-2 |
発行日 |
2018-06-05 (ITS) |
ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
PDFダウンロード |
ITS2018-2 |
|