A Vision-Based Hierarchical Framework for Autonomous Front-Vehicle Taillights Detection and Signal Recognition
Journal
IEEE Conference on Intelligent Transportation Systems
Journal Volume
2015-October
Pages
931-937
Date Issued
2015
Author(s)
Abstract
Automatically recognizing rear light signals of front vehicles can significantly improve driving safety by automatic alarm and taking actions proactively to prevent rear-end collisions and accidents. Much previous research only focuses on detecting brake signals at night. In this paper, we present the design and implementation of a robust hierarchical framework for detecting taillights of vehicles and estimating alert signals (turning and braking) in the daytime. The three-layer structure of the vision-based framework can obviously reduce both false positives and false negatives of taillight detection. Comparing to other existing work addressing nighttime detection, the proposed method is capable of recognizing taillight signals under different illumination circumstances. By carrying out contrast experiments with existing state-of-the-art methods, the results show the high detection rate of the framework in different weather conditions during the daytime. © 2015 IEEE.
SDGs
Other Subjects
Accidents; Intelligent systems; Intelligent vehicle highway systems; Signal processing; Transportation; Vehicles; Contrast experiment; Design and implementations; High detection rate; Rear-end collisions; Signal recognition; State-of-the-art methods; Taillight detection; Three-layer structures; Signal detection
Type
conference paper