Road infrared target detection with I-YOLO

Mingyuan Sun; Haochun Zhang; Ziliang Huang; Yueqi Luo; Yiyi Li

doi:10.1049/ipr2.12331

Road infrared target detection with I-YOLO

Mingyuan Sun
, Haochun Zhang^*
, Ziliang Huang
, Yueqi Luo
, Yiyi Li

^*Corresponding author for this work

School of Energy Science and Engineering, Harbin Institute of Technology
SAIC

Research output: Contribution to journal › Article › peer-review

31 Scopus citations

Abstract

The detection of road infrared targets is essential for autonomous driving. Different from RGB images, the acquisition of infrared images is unaffected by visible light. However, the signal-to-noise ratio still presents significant challenges. This study demonstrates an improved infrared target detection model for road infrared target detection. An advanced EfficientNet is incorporated to replace the conventional structure and enhance feature extraction. A Dilated-Residual U-Net is also introduced to reduce the noise of infrared images. Meanwhile, the k-means algorithm and data enhancement are implemented to improve the detection performance. The experimental results show that the mean average precision of the proposed model is observed to be 0.89 for the infrared road dataset with an average detection speed of 10.65 s⁻¹.

Original language	English
Pages (from-to)	92-101
Number of pages	10
Journal	IET Image Processing
Volume	16
Issue number	1
DOIs	https://doi.org/10.1049/ipr2.12331
State	Published - Jan 2022
Externally published	Yes

Keywords

deep learning
image processing
infrared image
target detection

Access to Document

10.1049/ipr2.12331

Cite this

@article{9100b7516d694462b1f823731ad9a751,

title = "Road infrared target detection with I-YOLO",

abstract = "The detection of road infrared targets is essential for autonomous driving. Different from RGB images, the acquisition of infrared images is unaffected by visible light. However, the signal-to-noise ratio still presents significant challenges. This study demonstrates an improved infrared target detection model for road infrared target detection. An advanced EfficientNet is incorporated to replace the conventional structure and enhance feature extraction. A Dilated-Residual U-Net is also introduced to reduce the noise of infrared images. Meanwhile, the k-means algorithm and data enhancement are implemented to improve the detection performance. The experimental results show that the mean average precision of the proposed model is observed to be 0.89 for the infrared road dataset with an average detection speed of 10.65 s−1.",

keywords = "deep learning, image processing, infrared image, target detection",

author = "Mingyuan Sun and Haochun Zhang and Ziliang Huang and Yueqi Luo and Yiyi Li",

note = "Publisher Copyright: {\textcopyright} 2021 The Authors. IET Image Processing published by John Wiley \& Sons Ltd on behalf of The Institution of Engineering and Technology.",

year = "2022",

month = jan,

doi = "10.1049/ipr2.12331",

language = "英语",

volume = "16",

pages = "92--101",

journal = "IET Image Processing",

issn = "1751-9659",

publisher = "John Wiley \& Sons Inc.",

number = "1",

}

TY - JOUR

T1 - Road infrared target detection with I-YOLO

AU - Sun, Mingyuan

AU - Zhang, Haochun

AU - Huang, Ziliang

AU - Luo, Yueqi

AU - Li, Yiyi

PY - 2022/1

Y1 - 2022/1

N2 - The detection of road infrared targets is essential for autonomous driving. Different from RGB images, the acquisition of infrared images is unaffected by visible light. However, the signal-to-noise ratio still presents significant challenges. This study demonstrates an improved infrared target detection model for road infrared target detection. An advanced EfficientNet is incorporated to replace the conventional structure and enhance feature extraction. A Dilated-Residual U-Net is also introduced to reduce the noise of infrared images. Meanwhile, the k-means algorithm and data enhancement are implemented to improve the detection performance. The experimental results show that the mean average precision of the proposed model is observed to be 0.89 for the infrared road dataset with an average detection speed of 10.65 s−1.

AB - The detection of road infrared targets is essential for autonomous driving. Different from RGB images, the acquisition of infrared images is unaffected by visible light. However, the signal-to-noise ratio still presents significant challenges. This study demonstrates an improved infrared target detection model for road infrared target detection. An advanced EfficientNet is incorporated to replace the conventional structure and enhance feature extraction. A Dilated-Residual U-Net is also introduced to reduce the noise of infrared images. Meanwhile, the k-means algorithm and data enhancement are implemented to improve the detection performance. The experimental results show that the mean average precision of the proposed model is observed to be 0.89 for the infrared road dataset with an average detection speed of 10.65 s−1.

KW - deep learning

KW - image processing

KW - infrared image

KW - target detection

UR - https://www.scopus.com/pages/publications/85114903518

U2 - 10.1049/ipr2.12331

DO - 10.1049/ipr2.12331

M3 - 文章

AN - SCOPUS:85114903518

SN - 1751-9659

VL - 16

SP - 92

EP - 101

JO - IET Image Processing

JF - IET Image Processing

IS - 1

ER -

Road infrared target detection with I-YOLO

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this