Vehicle detection in foggy weather combining millimeter wave rada and machine vision

LI Qi; YE Xiao-min; FENG Wen-bin

doi:10.37188/CJLCD.2022-0412

您当前的位置：

首页 >

文章列表页 >

Vehicle detection in foggy weather combining millimeter wave rada and machine vision

Image Processing | 更新时间：2023-10-10

- Vehicle detection in foggy weather combining millimeter wave rada and machine vision
- Chinese Journal of Liquid Crystals and Displays Vol. 38, Issue 10, Pages: 1445-1454(2023)
- 作者机构：
  
  1.陕西科技大学电子信息与人工智能学院，陕西西安 710021
- 作者简介：
- 基金信息：
  
  Project of Science and Technology Department of Shaanxi Province(S2023-YF-YBNY-0232);Xi'an Science and Technology Plan(22NYYF064);Xi'an Weiyang District Science and Technology Plan(201305)
- DOI：10.37188/CJLCD.2022-0412
  CLC：
扫描看全文
LI Qi, YE Xiao-min, FENG Wen-bin. Vehicle detection in foggy weather combining millimeter wave rada and machine vision. [J]. Chinese Journal of Liquid Crystals and Displays 38(10):1445-1454(2023)
DOI：

LI Qi, YE Xiao-min, FENG Wen-bin. Vehicle detection in foggy weather combining millimeter wave rada and machine vision. [J]. Chinese Journal of Liquid Crystals and Displays 38(10):1445-1454(2023) DOI： 10.37188/CJLCD.2022-0412.

摘要

车辆检测对于辅助驾驶系统至关重要，由于雾天道路场景严重退化，图像中的车辆信息不明显，导致车辆检测存在漏检、误检的问题。针对上述问题，本文提出了一种融合毫米波雷达和机器视觉的雾天车辆检测方法。首先，采用暗通道去雾算法对图像进行预处理，提高雾天图像中车辆信息的显著性。然后，采用知识蒸馏改进YOLOv5s算法，在YOLOv5s的特征提取网络中引入知识蒸馏，在目标定位和分类阶段计算蒸馏损失，对损失进行反向传播训练小型网络模型，在保证视觉检测准确度的同时提高检测速度。最后，采用基于潜在目标检测区域搜索的距离匹配算法对视觉检测结果和毫米波雷达检测结果进行决策级融合。以检测目标的类型和距离为匹配依据，滤除干扰信息和错误信息，保留毫米波雷达检测和视觉检测融合后的检测置信度较高的目标，从而提高车辆检测的准确率。实验结果表明，该方法在雾天下最高检测准确率达92.8%，召回率达90.7%，能够实现雾天对车辆的检测。

Abstract

Vehicle detection is very vital to the assisted driving system. Due to the serious degradation of the foggy road scene， the vehicle information in the image is not obvious， resulting in missed detection and false detection problems in vehicle detection. Aiming at the above problems， a vehicle detection method in foggy weather combining millimeter-wave radar and machine vision is proposed. First， the dark channel dehazing algorithm is used to preprocess the image to improve the salience of vehicle information in the image under foggy conditions. Then， the knowledge distillation is used to improve the YOLOv5s algorithm， and the knowledge distillation is introduced into the feature extraction network of YOLOv5s， which is used in the target positioning and classification stages to calculate the distillation loss and backpropagate the loss to train a small network model to improve the detection speed while ensuring the accuracy of visual detection. Finally， the distance matching algorithm based on the search of potential target detection areas is used to compare the visual detection results and the millimeter-wave radar detection results decision-making fusion. Based on the type and distance of the detected target， the interference information and erroneous information is filtered out， and the targets with high confidence after fusion in millimeter-wave radar detection and visual detection is retained. Thereby， the accuracy of vehicle detection is improved. The experimental results show that the method has the highest detection accuracy rate of 92.8% and the recall rate of 90.7% in foggy weather， which can realize the detection of vehicles in foggy weather.

关键词

车辆检测毫米波雷达去雾知识蒸馏距离匹配决策级融合

Keywords

vehicle detectionmillimeter wave radardefoggingknowledge distillationdistance matchingdecision level fusion

references

GUO Y， LIANG R L， CUI Y K， et al. A domain-adaptive method with cycle perceptual consistency adversarial networks for vehicle target detection in foggy weather ［J］. IET Intelligent Transport Systems， 2022， 16（7）： 971-981. doi: 10.1049/itr2.12190http://dx.doi.org/10.1049/itr2.12190

HE K M， SUN J， TANG X O， et al. Single image haze removal using dark channel prior ［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2011， 33（12）： 2341-2353. doi: 10.1109/tpami.2010.168http://dx.doi.org/10.1109/tpami.2010.168

GAO T， LI K， CHEN T， et al. A novel UAV sensing image defogging method ［J］. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing， 2020， 13： 2610-2625. doi: 10.1109/jstars.2020.2998517http://dx.doi.org/10.1109/jstars.2020.2998517

张炳力，詹叶辉，潘大巍，等. 基于毫米波雷达和机器视觉融合的车辆检测［J］. 汽车工程，2021，43（4）：478-484． doi: 10.19562/j.chinasae.qcgc.2021.04.004http://dx.doi.org/10.19562/j.chinasae.qcgc.2021.04.004

ZHANG B L， ZHAN Y H， PAN D W， et al. Vehicle detection based on fusion of millimeter-wave radar and machine vision ［J］. Automotive Engineering， 2021， 43（4）： 478-484. （in Chinese）. doi: 10.19562/j.chinasae.qcgc.2021.04.004http://dx.doi.org/10.19562/j.chinasae.qcgc.2021.04.004

JIANG Q Y， ZHANG L J， MENG D J. Target detection algorithm based on MMW radar and camera fusion ［C］//Proceedings of 2019 IEEE Intelligent Transportation Systems Conference. Auckland： IEEE， 2019： 1-6. doi: 10.1109/itsc.2019.8917504http://dx.doi.org/10.1109/itsc.2019.8917504

HAN S Y， WANG X， XU L H， et al. Frontal object perception for intelligent vehicles based on radar and camera fusion ［C］//Proceedings of the 35th Chinese Control Conference. Chengdu： IEEE， 2016： 4003-4008. doi: 10.1109/chicc.2016.7553978http://dx.doi.org/10.1109/chicc.2016.7553978

LUO Y， ZHAO Y F， LI J X， et al. Computational imaging without a computer： seeing through random diffusers at the speed of light ［J］. eLight， 2022， 2（1）： 4. doi: 10.1186/s43593-022-00012-4http://dx.doi.org/10.1186/s43593-022-00012-4

ZUO C， QIAN J M， FENG S J， et al. Deep learning in optical metrology： a review ［J］. Light： Science & Applications， 2022， 11（1）： 39. doi: 10.1038/s41377-022-00714-xhttp://dx.doi.org/10.1038/s41377-022-00714-x

Guohai SITU G. Deep holography ［J］. Light： Advanced Manufacturing， 2022， 3（2）： 278-300. doi: 10.37188/lam.2022.013http://dx.doi.org/10.37188/lam.2022.013

LIU H Y， SUN F Q， GU J， et al. SF-YOLOv5： a lightweight small object detection algorithm based on improved feature fusion mode ［J］. Sensors， 2022， 22（15）： 5817. doi: 10.3390/s22155817http://dx.doi.org/10.3390/s22155817

刘元峰，姬海军，刘立波. 基于YOLOv5s的高速公路车辆实时检测模型［J］. 液晶与显示，2022，37（9）：1228-1241． doi: 10.37188/CJLCD.2022-0026http://dx.doi.org/10.37188/CJLCD.2022-0026

LIU Y F， JI H J， LIU L B. Real-time detection model of highway vehicle based on YOLOv5s ［J］. Chinese Journal of Liquid Crystals and Displays， 2022， 37（9）： 1228-1241. （in Chinese）. doi: 10.37188/CJLCD.2022-0026http://dx.doi.org/10.37188/CJLCD.2022-0026

DAI G W， FAN J C. An industrial-grade solution for crop disease image detection tasks ［J］. Frontiers in Plant Science， 2022， 13： 921057. doi: 10.3389/fpls.2022.921057http://dx.doi.org/10.3389/fpls.2022.921057

XU D P， LIU Y Q， WANG Q， et al. Target detection based on improved hausdorff distance matching algorithm for millimeter-wave radar and video fusion ［J］. Sensors， 2022， 22（12）： 4562. doi: 10.3390/s22124562http://dx.doi.org/10.3390/s22124562

CHEN B， PEI X F， CHEN Z F. Research on target detection based on distributed track fusion for intelligent vehicles ［J］. Sensors， 2020， 20（1）： 56. doi: 10.3390/s20010056http://dx.doi.org/10.3390/s20010056

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

FRKDNet： feature refine semantic segmentation network based on knowledge distillation

Vehicle detection based on real-time traffic condition and adaptive pixel segmentation

Improved YOLOx-based vehicle detection method for low light environment

Pointpillars point cloud detection network based on knowledge distillation and location guidance

Related Author

No data

Related Institution

Guiyang Aluminum-magnesium Design and Research Institute Co.Ltd.

College of Big Data and Information Engineering， Guizhou University

School of Automation, Nanjing University of Information Science & Technology

Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology

Wuxi Research Institute, Nanjing University of Information Science & Techaologty

Address：No.3888 Dong Nanhu Road, Changchun, Jilin, China 130033 Postal code：130033
Tel：0431-86176059 Email：yjxs@ciomp.ac.cn
Technical support is provided by Beijing Founder electronics co., LTD 吉ICP备11002662号-17 京公网安备11010802024621
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰