Chinese Journal of Liquid Crystals and Displays, Volume. 38, Issue 10, 1445(2023)

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

Qi LI, Xiao-min YE*, and Wen-bin FENG
Author Affiliations
  • College of Electronic Information and Artificial Intelligence,Shaanxi University of Science and Technology,Xi'an 710021,China
  • show less
    Figures & Tables(16)
    Comparison of defogging effects
    Flowchart of dark channel prior defogging algorithm
    Comparison of results before and after defogging
    Distillation process diagram
    Comparison of YOLOv5s detection results before and after improvement
    mm-wave radar and vision sensor coordinate system
    Comparison of detection results before and after defogging
    Results of fusion detection of actual collected data
    Detection results of nuScenes dataset fusion
    • Table 1. Performance comparison of defogging algorithms

      View table
      View in Article

      Table 1. Performance comparison of defogging algorithms

      算法优点缺点
      自适应直方图均衡化使用范围广增强效果有限,鲁棒性差
      暗通道去雾稳定性高,实时性高去雾效果偏暗
      DehazeNet去雾效果好网络结构复杂,实时性差
    • Table 2. Comprehensive performance of five versions of YOLOv5 on the COCO dataset

      View table
      View in Article

      Table 2. Comprehensive performance of five versions of YOLOv5 on the COCO dataset

      模型mAP时间/ms参数/MFLOPs/Billion
      YOLOv5n28.0451.94.5
      YOLOv5s37.4987.216.5
      YOLOv5m45.422421.249.0
      YOLOv5l4943046.5109.1
      YOLOv5x50.776686.7205.7
    • Table 3. Comparison of test results

      View table
      View in Article

      Table 3. Comparison of test results

      检测网络准确率/%召回率/%预测时间/s模型参数/MB
      YOLOv5s92.6393.310.1713.70
      改进后YOLOv5s92.5693.310.119.85
    • Table 4. Correspondence between target cross-sections and target types

      View table
      View in Article

      Table 4. Correspondence between target cross-sections and target types

      类型公交车自行车汽车卡车
      RCS160~1802100~150200~300
    • Table 5. Object detection results before and after defogging

      View table
      View in Article

      Table 5. Object detection results before and after defogging

      天气原始图像去雾后图像
      准确率/%召回率/%准确率/%召回率/%
      薄雾85.8975.6991.0185.55
      大雾69.5265.4682.4571.43
      浓雾62.3444.4267.8947.62
    • Table 6. Statistics of target detection results

      View table
      View in Article

      Table 6. Statistics of target detection results

      天气情况车辆总数去雾后视觉检测毫米波雷达检测融合后检测
      准确率/%召回率/%准确率/%召回率/%准确率/%召回率/%
      薄雾1 52191.0185.5583.6388.1292.8190.70
      大雾87582.4571.4382.9786.8487.3688.81
      浓雾55667.8947.6282.5786.6584.7388.17
    • Table 7. Test results of nuScenes dataset

      View table
      View in Article

      Table 7. Test results of nuScenes dataset

      天气情况准确率/%召回率/%
      薄雾92.0190.37
      浓雾88.8789.03
      大雾84.7186.79
    Tools

    Get Citation

    Copy Citation Text

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

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Research Articles

    Received: Dec. 8, 2022

    Accepted: --

    Published Online: Oct. 25, 2023

    The Author Email: Xiao-min YE (1946175704@qq.com)

    DOI:10.37188/CJLCD.2022-0412

    Topics