Indoor intelligent driving scene-based traffic signal data set. Issue 1 (January 2021)

• Record Type: Journal Article
• Authors: Fu, En; Pan, Weiguo; Pan, Feng; Yang, Qing
• Is Part Of: Journal of physics. Volume 1774:Issue 1(2021)
• Publication Date: 2021-01
• Subjects: Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/1774/1/012045
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: In recent years, with the rapid development of intelligent driving technology, vision based traffic signal target detection is one of the urgent problems to be solved; especially with the extensive application of deep learning technology, the traffic signal data set is more and more important. Due to the variety of traffic signals, there is no data set that can cover all categories at home and abroad. This paper presents a traffic signal data set based on indoor traffic scene. The currently disclosed data category includes 22 categories, including traffic signs and traffic lights. Simultaneously, this paper also uses a two-stage deep learning method to conduct experiments. In order to facilitate other researchers to conduct further research on the disclosed data set in this paper.