Acta Optica Sinica, Volume. 38, Issue 7, 0712006(2018)
Infrared Ship Target Detection Method Based on Deep Convolution Neural Network
Figures & Tables(11)
Fig. 6. Flow chart of infrared ship target implementation with marker-controlled watershed and improved AlexNet
Fig. 8. Effect pictures of infrared ship targets by marker-controlled watershed segmentation
Fig. 9. Segmentation time comparison between marker-controlled watershed method and Selective Search method
Table 1. Structure of improved AlexNet
View tableTable 1. Structure of improved AlexNet
Layer Net Input Convolution size Stride Padding Output 1 Convolution 224×224×2 11×11×96 2 Same 112×112×96 2 Pool 112×112×96 - 2 - 56×56×96 3 Convolution 56×56×96 5×5×256 1 Same 56×56×256 4 Pool 56×56×256 - 2 - 28×28×256 5 Convolution 28×28×256 3×3×384 1 Same 28×28×384 6 Convolution 28×28×384 3×3×384 1 Same 28×28×384 7 Convolution 28×28×384 3×3×256 1 Same 28×28×256 8 Pool 28×28×256 - 2 - 14×14×256 9 Reorg 4 — 2 — 14×14×1024 10 Route 9, 8 — — — 14×14×1280 11 Convolution 14×14×1280 3×3 1 - 14×14×1280 12 FC 14×14×1280 — 4096 13 FC 4096 — 4096 14 FC — — 2
Tools
Get Citation
Copy Citation Text
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
Paper Information
Category: Instrumentation, Measurement and Metrology
Received: Feb. 6, 2018
Accepted: --
Published Online: Sep. 5, 2018
The Author Email: Dong Feng (dongfeng@mail.sit)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20