Fast lane detection based on deep convolutional neural network and automatic training data labeling

研究成果: Article

抄録

Lane detection or road detection is one of the key features of autonomous driving. In computer vision area, it is still a very challenging target since there are various types of road scenarios which require a very high robustness of the algorithm. And considering the rather high speed of the vehicles, high efficiency is also a very important requirement for practicable application of autonomous driving. In this paper, we propose a deep convolution neural network based lane detection method, which consider the lane detection task as a pixel level segmentation of the lane markings. We also propose an automatic training data generating method, which can significantly reduce the effort of the training phase. Experiment proves that our method can achieve high accuracy for various road scenes in real-time.

元の言語English
ページ(範囲)566-575
ページ数10
ジャーナルIEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
E102A
発行部数3
DOI
出版物ステータスPublished - 2019 3 1

Fingerprint

Lane Detection
Convolution
Labeling
Computer vision
Pixels
Neural Networks
Neural networks
Experiments
Computer Vision
High Efficiency
High Accuracy
High Speed
Segmentation
Pixel
Robustness
Real-time
Scenarios
Target
Requirements
Experiment

ASJC Scopus subject areas

  • Signal Processing
  • Computer Graphics and Computer-Aided Design
  • Electrical and Electronic Engineering
  • Applied Mathematics

これを引用

@article{1a9056a255034f339cfad712a72712bf,
title = "Fast lane detection based on deep convolutional neural network and automatic training data labeling",
abstract = "Lane detection or road detection is one of the key features of autonomous driving. In computer vision area, it is still a very challenging target since there are various types of road scenarios which require a very high robustness of the algorithm. And considering the rather high speed of the vehicles, high efficiency is also a very important requirement for practicable application of autonomous driving. In this paper, we propose a deep convolution neural network based lane detection method, which consider the lane detection task as a pixel level segmentation of the lane markings. We also propose an automatic training data generating method, which can significantly reduce the effort of the training phase. Experiment proves that our method can achieve high accuracy for various road scenes in real-time.",
keywords = "Automatic labeling, Deep neural network, Real-time lane detection",
author = "Xun Pan and Harutoshi Ogai",
year = "2019",
month = "3",
day = "1",
doi = "10.1587/transfun.E102.A.566",
language = "English",
volume = "E102A",
pages = "566--575",
journal = "IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences",
issn = "0916-8508",
publisher = "Maruzen Co., Ltd/Maruzen Kabushikikaisha",
number = "3",

}

TY - JOUR

T1 - Fast lane detection based on deep convolutional neural network and automatic training data labeling

AU - Pan, Xun

AU - Ogai, Harutoshi

PY - 2019/3/1

Y1 - 2019/3/1

N2 - Lane detection or road detection is one of the key features of autonomous driving. In computer vision area, it is still a very challenging target since there are various types of road scenarios which require a very high robustness of the algorithm. And considering the rather high speed of the vehicles, high efficiency is also a very important requirement for practicable application of autonomous driving. In this paper, we propose a deep convolution neural network based lane detection method, which consider the lane detection task as a pixel level segmentation of the lane markings. We also propose an automatic training data generating method, which can significantly reduce the effort of the training phase. Experiment proves that our method can achieve high accuracy for various road scenes in real-time.

AB - Lane detection or road detection is one of the key features of autonomous driving. In computer vision area, it is still a very challenging target since there are various types of road scenarios which require a very high robustness of the algorithm. And considering the rather high speed of the vehicles, high efficiency is also a very important requirement for practicable application of autonomous driving. In this paper, we propose a deep convolution neural network based lane detection method, which consider the lane detection task as a pixel level segmentation of the lane markings. We also propose an automatic training data generating method, which can significantly reduce the effort of the training phase. Experiment proves that our method can achieve high accuracy for various road scenes in real-time.

KW - Automatic labeling

KW - Deep neural network

KW - Real-time lane detection

UR - http://www.scopus.com/inward/record.url?scp=85064003273&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85064003273&partnerID=8YFLogxK

U2 - 10.1587/transfun.E102.A.566

DO - 10.1587/transfun.E102.A.566

M3 - Article

AN - SCOPUS:85064003273

VL - E102A

SP - 566

EP - 575

JO - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

JF - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

SN - 0916-8508

IS - 3

ER -