Acta Optica Sinica, Volume. 41, Issue 16, 1611004(2021)

Polarimetric Imaging Target Classification Method Based on Attention Mechanism

Rui Sun1,2,3, Xiaobing Sun1,3、*, Xiao Liu1,3, and Qiang Song1,2,3
Author Affiliations
  • 1Anhui Institute of Optics and Fine Mechanics, Hefei Institutes of Physical Science, Chinese Academy of Sciences, Hefei, Anhui 230031, China
  • 2University of Science and Technology of China, Hefei, Anhui 230026, China
  • 3Key Laboratory of Optical Calibration and Characterization, Chinese Academy of Sciences, Hefei, Anhui 230031, China
  • show less
    Figures & Tables(9)
    Characteristics of polarimetric imaging targets. (a1)(a2) S0, S1 and S2 samples of different iron flakes; (b1)(b2) S0, S1 and S2 samples of different bullets
    Flow charts of data processing. (a) Flow chart of data processing in CBAM; (b) flow chart of data processing in channel attention module; (c) flow chart of data processing in spatial attention module
    Flow charts of data processing in different modules. (a) Flow chart of data processing in convolution module; (b) flow chart of data processing in residual module
    Structural diagram of polarimetric imaging target classification network based on attention mechanism
    Experimental equipment and experimental environment. (a) SALSA polarization camera system; (b) outdoor shooting; (c) targets placed in grassland; (d) cropped target sample
    Experimental results for target classification in sandy area. (a) Classification accuracies of proposed method and traditional methods; (b) training losses of proposed method and traditional methods
    Experimental results for target classification in bare soil area. (a) Classification accuracies of proposed method and traditional methods; (b) training losses of proposed method and traditional methods
    Experimental results for target classification in grassland area. (a) Classification accuracie of proposed method and traditional methods; (b) training losses of proposed method and traditional methods
    • Table 1. Numbers of model parameters and running time

      View table

      Table 1. Numbers of model parameters and running time

      NetworkNumber of modelparameters /107Running time insandy area /sRunning time inbare soil area /sRunning time ingrassland area /s
      Resnet181.1181891192177197
      CBAM-Resnet181.1292276241221250
    Tools

    Get Citation

    Copy Citation Text

    Rui Sun, Xiaobing Sun, Xiao Liu, Qiang Song. Polarimetric Imaging Target Classification Method Based on Attention Mechanism[J]. Acta Optica Sinica, 2021, 41(16): 1611004

    Download Citation

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

    Category: Imaging Systems

    Received: Jan. 19, 2021

    Accepted: Mar. 18, 2021

    Published Online: Aug. 12, 2021

    The Author Email: Sun Xiaobing (xbsun@aiofm.ac.cn)

    DOI:10.3788/AOS202141.1611004

    Topics