Open Access
 Issue JNWPU Volume 37, Number 1, February 2019 35 - 40 https://doi.org/10.1051/jnwpu/20193710035 03 April 2019

© 2019 Journal of Northwestern Polytechnical University

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

## 1 基于全卷积网络的海天线检测

### 1.1 全卷积神经网络

 图1全卷积神经网络模型

VGGNET-16层次结构

 图2上采样过程

G为求出的梯度图像。

### 1.3 霍夫变换

1) 离散化极坐标空间;

2) 对图像任意边缘点进行坐标变换, 并将其参数离散化, 然后判断与哪个数组元素相对应, 并让该元素数组加一;

3) 比较参数数组元素累加结果, 最大值(ρi, θi)即为对应的直线即为检测的结果, 该直线对应的方程为

## 3 仿真分析

### 3.1 海天线检测仿真实验

 图3海天线检测实例1
 图4海天线检测实例2

### 3.2 舰船检测仿真实验

 图5舰船检测实例

## 4 结论

1) 本文提出的基于全卷积网络的海天线检测方法可以较为准确地对海天区域进行分割，从而检测出海天线的位置。在天空存在云层干扰时，OTSU由于分割效果不好导致海天线检测失败；在海杂波较强时，海面处行均值梯度值会超过海天线位置处，行均值梯度法检测的海天线偏向海洋，但本方法却可以适用于各种复杂的海天背景，鲁棒性更强。

2) 本文提出的基于四向梯度的舰船检测方法利用了目标的多梯度特性可以有效分割目标，并且利用目标区域的面积与方差特性进行筛选，从而降低了虚警率，本方法比基于中值滤波和Top-hat形态学的检测方法检测精度与召回率更高。

## References

1. Liu Shijun. Research on Infrared Ship Target Detection[D]. Chengdu, University of Electronic Science and Technology of China, 2011 (in Chinese) [Google Scholar]
2. An Bowen, Hu Chunnuan, Liu Jie, et al. Study of Sea-Sky-Line Detection Algorithm[J]. InfrareTechnology, 2015, 37(3): 196-199 (in Chinese) [Article] [Google Scholar]
3. Li Cuihong. Automatic Detection Method of Infrared Ship Target under Complex Sea-Sky Background[D]. Hunan, Hunan Normal University, 2012 (in Chinese) [Google Scholar]
4. Zou Ruibin, Shi Caicheng, Mao Erke. Shearlet-Based Infrared Target Detection Algorithm on Complex Sea[J]. Chinese Journal of Scientific Instrument, 2011, 32(5): 1104-1107 (in Chinese) [Article] [Google Scholar]
5. Hou Wang, Sun Xiaoliang, Shang Yang,et al. Present State and Perspectives of Small Infrared Targets Detection Technology[J]. Infrared Technology, 2015, 37(1): 1-10 (in Chinese) [Article] [Google Scholar]
6. Ding Peng, Zhang Ye, Liu Rang, et al. Infrared Small Target Detection Based on Adaptive Canny Algorithm and Morphology[J]. Chinese Journal of Liquid Crystals and Displays, 2016, 31(8): 793-800 (in Chinese) [Article] [CrossRef] [Google Scholar]
7. Evan Shelhamer, Jonathan Long, Trevor Darrell. Fully Convolutional Networks for Semantic Segmentation[J]. IEEE Trans on Pattern Analysis and Machine Intelligence, 2017, 39(4): 640-651 [Article] [CrossRef] [Google Scholar]
8. Vijay Badrinarayanan, Alex Kendall, Roberto Cipolla. SegNet:a Deep Convolutional Encoder-Decoder Architecture for Image Segmentation[J]. IEEE Trans on Pattern Analysis and Machine Intelligence, 2017, 39(12): 2481-2495 [Article] [Google Scholar]
9. Li Yanjun, Zhang Ke. Scene Matching and Target Recognition Technology[M]. Xi'an, Northwestern Polytechnical University Press, 2009 [Google Scholar]
10. Liu Shuai, Wei Xianzhi, Gao Xiaomei. A Target Detection Method Based on Median Filter and Multi-Grad Search Algorithm[J]. Electronics Optics and Control, 2011, 18(2): 81-84 (in Chinese) [Article] [Google Scholar]

VGGNET-16层次结构

## All Figures

 图1全卷积神经网络模型 In the text
 图2上采样过程 In the text
 图3海天线检测实例1 In the text
 图4海天线检测实例2 In the text
 图5舰船检测实例 In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.