Laser & Optoelectronics Progress, Volume. 61, Issue 24, 2412001(2024)

Vehicle Size Measurement and Information Identification Using an Improved DenseNet Approach

Shuanfeng Zhao*, Jian Yao, and Jia Li
Author Affiliations
  • College of Mechanical Engineering, Xi'an University of Science and Technology, Xi'an 710054, Shaanxi , China
  • show less
    Figures & Tables(17)
    Flow chart of vehicle size measurement and information identification
    CEDN structure
    DenseNet structure
    Schematic diagram of depthwise separable convolution
    Improved DenseNet structure diagram
    Overview of field experiment. (a) Schematic diagram of experimental device; (b) camera shooting scene
    Vehicle contour result diagrams
    Dataset display of buses and two-axle trucks
    Comparison of dimensional errors of different network models in predicting the same vehicle
    Different prediction results of the improved model for the same vehicle type
    • Table 1. Length, width, and height of different styles of Cadillac CT6

      View table

      Table 1. Length, width, and height of different styles of Cadillac CT6

      Vehicle modelLength /mmWidth /mmHeight /mm
      Cadillac CT6 2023522318901480
      Cadillac CT6 2022522318791498
      Cadillac CT6 2017517918791500
    • Table 2. Detailed parameters of buses and two-axle trucks

      View table

      Table 2. Detailed parameters of buses and two-axle trucks

      ManufacturerUseSpecific modelNumber of axlesLength /mmWidth /mmHeight /mm
      YutongbusZK6808BEVQ12804523603260
      Shaanxi automobilevan truckSX5049XXYBEV331L2599523003130
      Farizon commercial vehiclevan truckDNC5072XXYBEVK12599523503320
    • Table 3. Comparison of different models on self-made data sets

      View table

      Table 3. Comparison of different models on self-made data sets

      MethodRMSE /mmMAE /mmFLOPs /106Parameter quantity /106
      ResNet191.43147.161434.6514.69
      MobileNet V1274.27243.7359.964.13
      GhostNet252.75206.8414.375.22
      DenseNet153.48125.562868.448.25
      Improved DenseNet98.1686.62186.785.36
    • Table 4. Vehicle length prediction value

      View table

      Table 4. Vehicle length prediction value

      Vehicle modelLength /mmPredicted value /mm

      Error /

      %

      BYD Qin47074754.11.0
      Audi Q546344709.91.8
      DENZA D952505181.71.3
      Yutong bus80457958.31.1
      Shaanxi automoobile van truck59956048.90.9
    • Table 5. Vehicle width prediction value

      View table

      Table 5. Vehicle width prediction value

      Vehicle modelLength /mmPredicted value /mm

      Error /

      %

      BYD Qin17701791.21.2
      Audi Q518931934.62.2
      DENZA D919601914.22.3
      Yutong bus23602421.42.6
      Shaanxi automoobile van truck23002247.32.3
    • Table 6. Vehicle height prediction value

      View table

      Table 6. Vehicle height prediction value

      Vehicle modelLength /mmPredicted value /mmError /%
      BYD Qin14871497.40.7
      Audi Q516491675.31.6
      DENZA D919201895.11.3
      Yutong bus32603299.121.2
      Shaanxi automoobile van truck31303155.040.8
    • Table 7. Tesla Model 3 parameter information

      View table

      Table 7. Tesla Model 3 parameter information

      Number of axlesWheelbase /mmFront suspensionRear suspensionTire
      22875Double wishbone independent suspensionMulti-link independent suspension235/45 R18
    Tools

    Get Citation

    Copy Citation Text

    Shuanfeng Zhao, Jian Yao, Jia Li. Vehicle Size Measurement and Information Identification Using an Improved DenseNet Approach[J]. Laser & Optoelectronics Progress, 2024, 61(24): 2412001

    Download Citation

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

    Category: Instrumentation, Measurement and Metrology

    Received: Mar. 5, 2024

    Accepted: Apr. 25, 2024

    Published Online: Dec. 19, 2024

    The Author Email:

    DOI:10.3788/LOP240825

    CSTR:32186.14.LOP240825

    Topics