Journals Articles Conferences News About CLP
Submit a paper Email Alert
Search
Search
Articles
Journals
News
Advanced Search
Top Searches
2D MaterialsTransformation opticsQuantum PhotonicsSmall lasersSemiconductor UV PhotonicsSupersymmetric microring laser arrays

Laser & Optoelectronics Progress, Volume. 59, Issue 11, 1106006(2022)

Real-Time Acquisition and Positioning Technology of Unmanned Aerial Vehicle Optical Communication Based on Improved YOLOv4 Network

Tingzuo Chen, Xiaolong Ni, Suping Bai*, and Xin Yu
Author Affiliations
  • School of Electro-Optical Engineering, Changchun University of Science and Technology, Changchun 130022, Jilin , China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (15)
    Equations (9)
    References (1)
    Tools
    Paper Information
    Topics
    Figures & Tables(15)
    Schematic diagram of the system

    Fig. 1. Schematic diagram of the system

    Download full size
    Structure of the system

    Fig. 2. Structure of the system

    Download full size
    Feature map attributes predicted by the YOLOv4 network

    Fig. 3. Feature map attributes predicted by the YOLOv4 network

    Download full size
    Simplified structure of the improved YOLOv4 network

    Fig. 4. Simplified structure of the improved YOLOv4 network

    Download full size
    Connection mode of three network structures. (a) Connection mode 1; (b) connection mode 2; (c) connection mode 3

    Fig. 5. Connection mode of three network structures. (a) Connection mode 1; (b) connection mode 2; (c) connection mode 3

    Download full size
    Principles of two feature fusion modes. (a) Concatenate; (b) add

    Fig. 6. Principles of two feature fusion modes. (a) Concatenate; (b) add

    Download full size
    Flow chart of the PID algorithm

    Fig. 7. Flow chart of the PID algorithm

    Download full size
    Some images in the data set

    Fig. 8. Some images in the data set

    Download full size
    Training results of the beacon spot data set

    Fig. 9. Training results of the beacon spot data set

    Download full size
    Training results of different networks. (a) Loss function; (b) mAP

    Fig. 10. Training results of different networks. (a) Loss function; (b) mAP

    Download full size
    Captured alignment result of the UAV. (a) Indoor environment; (b) background glare interference environment; (c) flight status

    Fig. 11. Captured alignment result of the UAV. (a) Indoor environment; (b) background glare interference environment; (c) flight status

    Download full size

    • Table 1. Recognition results of the test set

      View table

      Table 1. Recognition results of the test set

      CountCategoryXTPXFPXFNXTN
      518beacon spot5032113

    • Table 2. Statistics of PASCAL VOC 2007 and PASCAL VOC 2012 data sets

      View table

      Table 2. Statistics of PASCAL VOC 2007 and PASCAL VOC 2012 data sets

      Data setTraining setValidation setTraining set +validation setTest setTotal data set
      ImageObjectImageObjectImageObjectImageObjectImageObject
      Total821819910833320148165514005816492394823304379540
      PASCAL VOC 20072501630125106207501112608495212032996324640
      PASCAL VOC 2012571713609582313841115402745011540274502308054900

    • Table 3. Detection results of different networks on the PASCAL VOC 2007 test set

      View table

      Table 3. Detection results of different networks on the PASCAL VOC 2007 test set

      NetworkTrainmAP /%FPS
      YOLOv4PASCAL VOC 2007+PASCAL VOC 2012768
      YOLOv4.tinyPASCAL VOC 2007+PASCAL VOC 20124947
      Improved YOLOv4PASCAL VOC 2007+PASCAL VOC 20126342

    • Table 4. Parameters of NVIDIA Jetson Xavier NX embedded system

      View table

      Table 4. Parameters of NVIDIA Jetson Xavier NX embedded system

      ParameterNVIDIA Jetson Xavier NX embedded system
      GPUNVIDIA Volta architecture with 384 NVIDIA CUDA® cores and 48 Tensor cores
      CPU6-core NVIDIA Carmel ARM®v8.2 64-bit CPU 6 MB L2 + 4 MB L3
      Memory8 GB 128-bit LPDDR4x @ 51.2 Gbit/s
      StoragemicroSD（128 G）
      Mechanical103 mm×90.5 mm×34.66 mm
    Tools

    Get Citation

    Copy Citation Text

    Tingzuo Chen, Xiaolong Ni, Suping Bai, Xin Yu. Real-Time Acquisition and Positioning Technology of Unmanned Aerial Vehicle Optical Communication Based on Improved YOLOv4 Network[J]. Laser & Optoelectronics Progress, 2022, 59(11): 1106006

    Download Citation

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

    Category: Fiber Optics and Optical Communications

    Received: Aug. 12, 2021

    Accepted: Aug. 31, 2021

    Published Online: Jun. 9, 2022

    The Author Email: Suping Bai (2541533443@qq.com)

    DOI:10.3788/LOP202259.1106006

    Topics

    laser devices and laser physics
    Lasers and Laser Optics
    Laser physics
    laser manufacturing
    Instrumentation, Measurement and Metrology