Acta Optica Sinica, Volume. 38, Issue 1, 0111005(2018)

Valid Aircraft Detection System for Remote Sensing Images Based on Cognitive Models

Yuqingyang Hou*, Jicheng Quan, and Yongming Wei
Author Affiliations
  • Laboratory of Digital Earth Science, Aviation University of Air Force, Changchun, Jilin 130000, China
  • show less
    Figures & Tables(17)
    Structural diagram of detection system
    Workflow chart of system
    Structural diagram of SSD network
    Structure of full convolution segmentation network
    Process diagram of filtering out invalid targets
    Training process of system
    Airport detection samples
    Detection samples for unshaded aircrafts
    Detection samples for shaded aircrafts
    Contrast between (a) artificial segmentation image and (b) original image
    Sketch of fIoU definition
    (a) Training accuracy and (b) loss parameter versus iteration number in classification training process on VGGNet
    Aircraft detection results
    Airport segmentation results
    Screening results from aircraft detection
    • Table 1. Detection results

      View table

      Table 1. Detection results

      TargetfAPAverage fIoUTime /(ms·frame-1)
      Airport76.610.64500
      Aircraft85.460.9510
      Shaded aricraft80.230.8913
    • Table 2. Performance contrast among different detection algorithms

      View table

      Table 2. Performance contrast among different detection algorithms

      AlgorithmfAPEfficient aircraft ratioTime /(ms·frame-1)
      DPM80.6154.2%670
      SSD40.1263.1%120
      Recognitive model75.2394.5%528
    Tools

    Get Citation

    Copy Citation Text

    Yuqingyang Hou, Jicheng Quan, Yongming Wei. Valid Aircraft Detection System for Remote Sensing Images Based on Cognitive Models[J]. Acta Optica Sinica, 2018, 38(1): 0111005

    Download Citation

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

    Category: Imaging Systems

    Received: Jun. 20, 2017

    Accepted: --

    Published Online: Aug. 31, 2018

    The Author Email: Hou Yuqingyang (894210081@qq.com)

    DOI:10.3788/AOS201838.0111005

    Topics