Detection of the Traffic Light in Challenging Environmental Conditions

EasyChair Preprint no. 4950

Abstract

This paper is to recognize the current traffic light phase to reliably detect the status of the traffic light. eg; Red, Green and, Off-status and to focus cases where cameras have difficulties eg; Traffic light high above the ground, Out of field-of-view, the traffic light is front-or backlit by the sun and to achieve an improved redundancy leading to batter confidence in driving assistance systems and for future automated driving solutions. In this paper,  two methods are used;  The first method presents a vision-based traffic light structure detection and convolutional neural network (CNN) based state recognition method, which is robust under different illumination and weather conditions. The second method presents a deep fusion network for robust fusion without a large corpus of labeled training data covering all asymmetric distortions.

Keyphrases: CNN, deep fusion networks, traffic light detection.

@Booklet{EasyChair:4950,
  author = {Syeda Shahista and Afreen Khatoon Khan},
  title = {Detection of the Traffic Light in Challenging Environmental Conditions},
  howpublished = {EasyChair Preprint no. 4950},

  year = {EasyChair, 2021}}