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

Acta Optica Sinica, Volume. 41, Issue 10, 1028001(2021)

Real-Time Object Detection in Remote Sensing Images Based on Embedded System

Yuanjun Nong and Junjie Wang*
Author Affiliations
  • School of Engineering, Ocean University of China, Qingdao, Shandong 266100, China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (12)
    Equations (0)
    References (19)
    Cited By (7)
    Tools
    Paper Information
    Topics
    Figures & Tables(12)
    Network structure of YOLOv3-tiny

    Fig. 1. Network structure of YOLOv3-tiny

    Download full size
    Detection flow of YOLOv3-tiny

    Fig. 2. Detection flow of YOLOv3-tiny

    Download full size
    52×52 scale shallow features and 26×26 scale deep features

    Fig. 3. 52×52 scale shallow features and 26×26 scale deep features

    Download full size
    Multi-scale prediction structure. (a) Multi-scale prediction before modified; (b) multi-scale prediction after modified

    Fig. 4. Multi-scale prediction structure. (a) Multi-scale prediction before modified; (b) multi-scale prediction after modified

    Download full size
    Spatial attention module. (a) Original spatial attention module; (b) modified spatial attention module

    Fig. 5. Spatial attention module. (a) Original spatial attention module; (b) modified spatial attention module

    Download full size
    Network structure of YOLO-RS

    Fig. 6. Network structure of YOLO-RS

    Download full size
    Loss curve during training process

    Fig. 7. Loss curve during training process

    Download full size
    Test results of YOLO-RS

    Fig. 8. Test results of YOLO-RS

    Download full size
    Experiments on the embedded platform Jetson Xavier NX

    Fig. 9. Experiments on the embedded platform Jetson Xavier NX

    Download full size
    Comparison of detection accuracy and speed of different models

    Fig. 10. Comparison of detection accuracy and speed of different models

    Download full size

    • Table 1. Detection performance of different methods on remote sensing test set

      View table

      Table 1. Detection performance of different methods on remote sensing test set

      MethodInput sizemAP /%Recall /%F1 /%BFLOPSVolume /MB
      YOLOv3608×60880.078181139.64246.5
      YOLOv3-tiny608×60873.09677211.6734.8
      YOLO-RS608×60876.7075787.0410.0
      YOLOv3512×51279.61768099.02246.5
      YOLOv3-tiny512×51271.6063708.2734.8
      YOLO-RS512×51275.5673774.9910.0
      YOLOv3416×41677.04707765.37246.5
      YOLOv3-tiny416×41665.1554645.4634.8
      YOLO-RS416×41672.8665723.3010.0

    • Table 2. Detection speed of different methods on Jetson Xavier NX

      View table

      Table 2. Detection speed of different methods on Jetson Xavier NX

      MethodSpeed /(frame·s-1)
      416×416512×512608×608
      YOLOv38.16.34.6
      YOLOv3-tiny54.640.831.5
      YOLO-RS56.743.532.5
    Tools

    Get Citation

    Copy Citation Text

    Yuanjun Nong, Junjie Wang. Real-Time Object Detection in Remote Sensing Images Based on Embedded System[J]. Acta Optica Sinica, 2021, 41(10): 1028001

    Download Citation

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

    Category: Remote Sensing and Sensors

    Received: Oct. 22, 2020

    Accepted: Dec. 30, 2020

    Published Online: May. 8, 2021

    The Author Email: Wang Junjie (wjj@ouc.edu.cn)

    DOI:10.3788/AOS202141.1028001

    Topics

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