Journal of South China University of Technology(Natural Science Edition) ›› 2022, Vol. 50 ›› Issue (3): 1-8.doi: 10.12141/j.issn.1000-565X.210352

Special Issue: 2022年交通运输工程

• Traffic & Transportation Engineering • Previous Articles     Next Articles

Weather Recognition of Highway Surveillance Scenes Based on Light-Weight Deep Neural Network

FU Xinsha1 ZENG Yanjie2 MA Li3 HU Hongyi1 HU Jiacheng4 TANG Feng1   

  1. 1.School of Civil Engineering and Transportation, South China University of Technology, Guangzhou 510640, Guangdong, China；
    2.  Guangdong Provincial Transport Planning and Research Center, Guangzhou 510101, Guangdong, China;
    3.Department of Mathematics and Computer Engineering, Ordos Institute of Technology, Ordos 017000, Inner Mongolia, China;
    4.  CCCC-FHDI Engineering Co., Ltd., Guangzhou 510000,Guangdong, China
  • Received:2021-06-01 Revised:2021-10-09 Online:2022-03-25 Published:2022-03-01
  • Contact: 马丽(1978-)，女，副教授，主要从事视频处理、深度学习和行为识别研究。 E-mail:lima@oit.edu.cn
  • About author:符锌砂(1955-)，男，教授，博士生导师，主要从事智能交通系统、道路规划与设计。E-mail：fuxinsha@163.com
  • Supported by:
    Supported by the National Natural Science Foundation of China（51978283,51778242）

PDF

335

Abstract: For the problem that deep convolutional neural network models are difficult to deploy at the edge due to their large number of parameters and high hardware device requirements, this paper researched on the weather re-cognition algorithm based on light-weight neural network with the application scenario of highway surveillance images. The light-weight neural network model MobileNet was first analyzed theoretically, and the difference between the deeply separable convolution operation and the standard convolution operation was analyzed in terms of the number of parameters and the number of computations. At the same time, a weather recognition dataset based on highway surveillance images was collected and labeled. Based on this, models including several light-weight neural networks were built and trained for comparison experiments, and the experimental results verified the advantages of MobileNet in terms of recognition accuracy, speed and number of model parameters. In addition, this paper explored the feature representation of MobileNet as well as the inter-class separability and intra-class clustering of features by the visualization algorithm t-SNE in terms of both class activation analysis and feature distribution, and the results further supported the above analysis.

Key words: light-weight deep neural network, weather recognition, highway surveillance images, feature distribution visualization

CLC Number: