Fast Obstacle Detection Platform Based on Yolo

Sie sind hier: Homepage > Suche

Fast Obstacle Detection Platform Based on Yolo

Ma, Wenhua / Wang, Xin / Wang, Baohua

Due to its energy saving and environmental protection, large passenger capacity, high speed, and high punctuality, subway transportation has received unanimous praise from the public. However, train operation accidents happen from time to time. In order to reduce the accident rate in train operation, people have put forward higher requirements for the safe and stable operation of trains. Due to the development of unmanned driving technology, a vision-based active obstacle detection system is urgently needed for the operation of the subway. In this experiment, the YOLO network (single step detection network) was used. This platform can effectively detect target obstacles in the image. The mAP value reaches 95.47%, the pedestrian accuracy rate reaches 94.75%, and the train accuracy rate reaches 96.19%. At the same time, how to identify the target in a specific area and shorten the time of image recognition is also the direction of future research.

Zugriff

Zugriff prüfen

Verfügbarkeit in meiner Bibliothek prüfen

Bestellung bei Subito €

Exportieren, teilen und zitieren

Dokumentinformationen

Titel :

Fast Obstacle Detection Platform Based on Yolo

Weitere Titelangaben:

Lect. Notes Electrical Eng.

Beteiligte:

Qin, Yong (Herausgeber:in) / Jia, Limin (Herausgeber:in) / Liang, Jianying (Herausgeber:in) / Liu, Zhigang (Herausgeber:in) / Diao, Lijun (Herausgeber:in) / An, Min (Herausgeber:in) / Ma, Wenhua (Autor:in) / Wang, Xin (Autor:in) / Wang, Baohua (Autor:in)

Kongress:

International Conference on Electrical and Information Technologies for Rail Transportation ; 2021 October 21, 2021 - October 23, 2021

Erschienen in:

Proceedings of the 5th International Conference on Electrical Engineering and Information Technologies for Rail Transportation (EITRT) 2021 ; Kapitel : 25 ; 218-226

Lecture Notes in Electrical Engineering ; 868

Erscheinungsdatum :

2022-02-23

Format / Umfang :

9 pages

ISBN :

978-981-16-9913-9 , 978-981-16-9912-2

ISSN :