汽车工程 ›› 2021, Vol. 43 ›› Issue (4): 485-491.doi: 10.19562/j.chinasae.qcgc.2021.04.005

• • 上一篇    下一篇

基于语义分割网络的路面积水与湿滑区域检测

王海1(),蔡柏湘1,蔡英凤2,刘泽2,孙恺3,陈龙2   

  1. 1.江苏大学汽车与交通工程学院,镇江 212013
    2.江苏大学汽车工程研究院,镇江 212013
    3.上海禾赛科技股份有限公司,上海 201702
  • 收稿日期:2020-09-28 修回日期:2021-01-24 出版日期:2021-04-25 发布日期:2021-04-23
  • 通讯作者: 王海 E-mail:wanghai1019@163.com
  • 基金资助:
    国家重点研发计划(2018YFB0105000);国家自然科学基金(U20A20333);江苏省自然科学基金(BK20180100);江苏省六大人才高峰项目(2018-TD-GDZB-022);江苏省重点研发项目(BE2019010-2)

Detection of Water⁃covered and Wet Areas on Road Pavement Based on Semantic Segmentation Network

Hai Wang1(),Baixiang Cai1,Yingfeng Cai2,Ze Liu2,Kai Sun3,Long Chen2   

  1. 1.School of Automotive and Traffic Engineering,Jiangsu University,Zhenjiang 212013
    2.Institute of Automotive Engineering,Jiangsu University,Zhenjiang 212013
    3.Hesai Instruments Inc. ,Shanghai 201702
  • Received:2020-09-28 Revised:2021-01-24 Online:2021-04-25 Published:2021-04-23
  • Contact: Hai Wang E-mail:wanghai1019@163.com

摘要:

积水或湿滑路面的道路附着系数远小于干燥路面的附着系数,对交通的安全性和机动性都有很大的影响。通过及时获取路面状态信息而发出预警,可大大减小潜在伤害。本文中研究了基于图像的语义分割网络在积水和潮湿的路面状态识别中的应用,它不仅可预测未来路面状态信息,且可得到路面积水和湿滑区域的分布。该方法利用语义分割网络Res-UNet++,分割出路面的积水和湿滑区域。Res-UNet++结构包括嵌套了不同深度的编码器-解码器结构,并在网络的特征提取部分加入残差结构,从而使图像的特征更容易学习。该方法取得了平均交并比为90.07%的分割精度并克服了其它方法的缺点。

关键词: 积水与湿滑区域检测, 编码器-解码器, 深度学习, 语义分割网络

Abstract:

The adhesion coefficient of water?covered or wet road surfaces is much smaller than that of dry road surfaces, which has a great effect on the traffic safety and maneuverability of vehicle, so by timely obtaining the information of road conditions and issuing warning can greatly reduce potential impairments. In this paper, the application of image?based semantic segmentation network to the recognition of water?covered and wet road conditions is studied, which can not only predict the road conditions in the future, but also obtain the distribution of water?covered and wet areas on road. The method proposed utilizes the semantic segmentation network Res-UNet++ to segment the water?covered and wet areas on road pavement. The Res-UNet++ structure includes an embedded encoder?decoder structure of different depths, with a residual structure added on the feature extraction part of the network to make image features easier to learn. The method adopted achieves an average segmentation accuracy of 90.07% in MIoU, and overcomes the defects of other methods.

Key words: detection of water?covered and wet areas, encoder?decoder, deep learning, semantic segmentation network