Traffic Information Detection

EasyChair Preprint 2337, version 1

Abstract

In recent years, with a number of technology breakthroughs in the world, one of the Artificial Intelligence (AI) branches self-driving vehicle is becoming closer to our lives. In this paper we built a model which contains two submodules: lane detection and vehicle detection. Our lane detection model is based on a heuristic approach to detect lanes. It can be broken down into three steps: Image preprocess, Lane edge points identification, and lane cure generation. As for the vehicle detection, we applied YOLO series algorithms which are fast, accurate and can be used in real-time detection.

Keyphrases: computer vision, deep learning, lane detection, vehicle detection

@booklet{EasyChair:2337,
  author    = {Donghao Qiao and Jiayuan Zhou and Farhana Zulkernine},
  title     = {Traffic Information Detection},
  howpublished = {EasyChair Preprint 2337},
  year      = {EasyChair, 2020}}