• Acta Optica Sinica
  • Vol. 38, Issue 7, 0712006 (2018)
Wenxiu Wang1、2、3, Yutian Fu1、2, Feng Dong1、2、*, and Feng Li1、2、3
Author Affiliations
  • 1 Key Laboratory of Infrared System Detection and Imaging Technology, Chinese Academy of Sciences, Shanghai 200083, China
  • 2 Shanghai Institute of Technical Physics, Chinese Academy of Sciences, Shanghai 200083, China
  • 3 University of Chinese Academy of Sciences, Beijing 100049, China
  • show less
    DOI: 10.3788/AOS201838.0712006 Cite this Article Set citation alerts
    Wenxiu Wang, Yutian Fu, Feng Dong, Feng Li. Infrared Ship Target Detection Method Based on Deep Convolution Neural Network[J]. Acta Optica Sinica, 2018, 38(7): 0712006 Copy Citation Text show less
    Acquisition step marker-controlled watershed segmentation image
    Fig. 1. Acquisition step marker-controlled watershed segmentation image
    Realization in detail
    Fig. 2. Realization in detail
    CNN diagram
    Fig. 3. CNN diagram
    Improved AlexNet
    Fig. 4. Improved AlexNet
    Reorg implementation
    Fig. 5. Reorg implementation
    Flow chart of infrared ship target implementation with marker-controlled watershed and improved AlexNet
    Fig. 6. Flow chart of infrared ship target implementation with marker-controlled watershed and improved AlexNet
    Two effect pictures by marker-controlled watershed segmentation
    Fig. 7. Two effect pictures by marker-controlled watershed segmentation
    Effect pictures of infrared ship targets by marker-controlled watershed segmentation
    Fig. 8. Effect pictures of infrared ship targets by marker-controlled watershed segmentation
    Segmentation time comparison between marker-controlled watershed method and Selective Search method
    Fig. 9. Segmentation time comparison between marker-controlled watershed method and Selective Search method
    Final recognition results
    Fig. 10. Final recognition results
    LayerNetInputConvolution sizeStridePaddingOutput
    1Convolution224×224×211×11×962Same112×112×96
    2Pool112×112×96-2-56×56×96
    3Convolution56×56×965×5×2561Same56×56×256
    4Pool56×56×256-2-28×28×256
    5Convolution28×28×2563×3×3841Same28×28×384
    6Convolution28×28×3843×3×3841Same28×28×384
    7Convolution28×28×3843×3×2561Same28×28×256
    8Pool28×28×256-2-14×14×256
    9Reorg4214×14×1024
    10Route9, 814×14×1280
    11Convolution14×14×12803×31-14×14×1280
    12FC14×14×12804096
    13FC40964096
    14FC2
    Table 1. Structure of improved AlexNet
    Wenxiu Wang, Yutian Fu, Feng Dong, Feng Li. Infrared Ship Target Detection Method Based on Deep Convolution Neural Network[J]. Acta Optica Sinica, 2018, 38(7): 0712006
    Download Citation