﻿ 关联数据分析的船舶航行图像目标检测系统
 舰船科学技术  2022, Vol. 44 Issue (9): 178-181    DOI: 10.3404/j.issn.1672-7649.2022.09.038 PDF

Design of ship navigation image target detection system based on correlation data analysis
XIAO Ying-nan, SUN Shu-yu
College Engineering and Technical, Chengdu University of Technology, Leshan 614000, China
Abstract: Marine ship target detection technology is widely used in enemy ship monitoring, target reconnaissance and other scenes in the military field. It is of great significance to improve the target detection quality of marine ship navigation image. The core of this paper is to analyze and process the ship navigation image based on the correlation data analysis technology. A grey correlation algorithm is adopted. The steps of sea land separation, image smoothing and image enhancement of ship navigation image are introduced in detail. Through the development of software and hardware, a ship navigation image target detection system based on Windows 10 platform is built, which has certain practical application value.
Key words: correlation data analysis     navigation image     target detection     smoothing     grey correlation algorithm
0 引　言

1 船舶航行图像目标检测系统的概要设计

 图 1 船舶航行图像目标检测系统的基本框架 Fig. 1 Basic framework of ship navigation image target detection system

1）检测与识别模块

2）输出模块

3）交互模块

2 基于关联数据分析技术的船舶航行图像目标检测 2.1 灰色关联分析理论

 图 2 灰色关联分析流程图 Fig. 2 Grey correlation analysis flow chart

1)确认系统中的参考序列和比较序列

 ${X_0} = \left( {{x_0}\left( 1 \right),{x_0}\left( 2 \right),\cdots,{x_0}\left( n \right)} \right) \text{，}$

 ${X_i} = \left( {{x_i}\left( 1 \right),{x_i}\left( 2 \right),\cdots,{x_i}\left( n \right)} \right)\;\;i = 1,2,\cdots,n 。$

2)无量纲化处理

 ${X_0}^* = \left( {{x_0}\left( 1 \right)^*,{x_0}\left( 2 \right)^*,\cdots,{x_0}\left( n \right)^*} \right) \text{，}$

 ${X_i}^* = \left( {{x_i}\left( 1 \right)^*,{x_i}\left( 2 \right)^*,\cdots,{x_i}\left( n \right)^*} \right)\;\;i = 1,2,\cdots,n \text{。}$

3)计算序列的灰色关联系数

 $\gamma \left( {X_0^*(k),X_i^*(k)} \right) = \frac{{m + \rho M}}{{{\Delta _i}(k) + \rho M}} \text{，}$

 ${\Delta _i}(k) = X_0^*(k) - X_i^*(k) 。$

4)计算灰色关联度

 $\gamma \left( {X_0^*,X_i^*} \right) = \frac{1}{n}\sum\limits_{i = 1}^n {} \gamma \left( {X_0^*(k),X_i^*(k)} \right) 。$

2.2 船舶航行图像的噪声抑制

 $P(i,j) = K\left\{ \begin{gathered} \left( {{x_1},{y_1}} \right),\left( {{x_2},{y_2}} \right) \in M \times N\mid \hfill \\ \varphi \left( {{x_1},{y_1}} \right) = i \hfill \\ \varphi \left( {{x_1},{y_2}} \right) = j \hfill \\ \end{gathered} \right\} \text{，}$

 $T(\varepsilon ,t) = \frac{1}{{\sqrt \varepsilon }}\int_{ - \infty }^\infty {P(i,j)f(t)\left( {\frac{{t - \tau }}{\varepsilon }} \right)} {\rm{d}}t \text{。}$

 $F(t) = \frac{1}{\displaystyle{\int_{ - \infty }^\infty {P(i,j){e^{ - jwt}}{\rm{d}}t} }}\int_{ - \infty }^\infty {\frac{{{\rm{d}}s}}{{{s^2}}}\int_{ - \infty }^\infty {T(\varepsilon ,t)} } 。$
2.3 基于关联数据分析的船舶航行图像海陆分离

1）确定船舶航行图像的灰度图[3]

2）计算图像的最大和最小灰度 ${Z_i}/{Z_k}$ ，确定图像分割的阈值初始值为 $T = \dfrac{{\left( {{Z_i} + {Z_k}} \right)}}{2}$

3）基于灰度关联分析理论，进行像素点的关联度分析。

4）根据阈值计算海陆两部分的平均灰度值，如下：

 $\left\{ {\begin{array}{*{20}{l}} {{Z_{sea}} = \dfrac{\displaystyle {\sum\limits_{z(i,j) < {T^{}}} z (i,j) \times N(i,j)}}{\displaystyle {\sum\limits_{z(i,j) < {T^{}}} N (i,j)}}}, \\ {{Z_{land}} = \dfrac{\displaystyle{\sum\limits_{z(i,j) > T} z (i,j) \times N(i,j)}}{\displaystyle {\sum\limits_{z(i,j) > T} N (i,j)}}} 。\end{array}} \right. \text{}$

5）不断迭代计算，获取船舶航行图像的海陆分离结果。

 图 3 目标图像海陆分离方法效果图 Fig. 3 Effect drawing of sea land separation method of target image
2.4 船舶航行图像的灰度平滑处理

 图 4 多项式平滑处理算法示意图 Fig. 4 Schematic diagram of polynomial smoothing algorithm

 $y = {a_0} + {a_1}x + {a_2}{x^2} + \ldots + {a_{k - 1}}{x^{k - 1}} \text{，}$

 $\left[ {\begin{array}{*{20}{c}} {{y_{ - m}}} \\ {{y_{ - m + 1}}} \\ \vdots \\ {{y_n}} \end{array}} \right] = \left[ {\begin{array}{*{20}{c}} 1&{ - m}& \cdots &{{{( - m)}^{k - 1}}}\\ 1&{ - m + 1}& \cdots &{{{( - m)}^{k - 1}}} \\ \vdots & \vdots & \vdots & \vdots \\ 1&n& \cdots &{{n^{k - 1}}} \end{array}} \right]\left[ {\begin{array}{*{20}{c}} {{a_0}} \\ {{a_1}} \\ \vdots \\ {{a_n}} \end{array}} \right]+ \left[ {\begin{array}{*{20}{c}} {{e_{ - m}}} \\ {{e_{ - m + 1}}} \\ \vdots \\ {{e_n}} \end{array}} \right] \text{，}$

2.5 基于关联数据分析的航行图像目标检测

1）图像滤波

2）灰度平滑处理

3）关联度分析

 $\gamma \left( {Z_0^*,Z_i^*} \right) = \frac{1}{n}\sum\limits_{i = 1}^n {} \gamma \left( {Z_0^*(k),Z_i^*(k)} \right) \text{。}$

4）图像增强

 ${\nabla ^2}f = \frac{{{\partial ^2}f}}{{\partial {x^2}}} + \frac{{{\partial ^2}f}}{{\partial {y^2}}} \text{。}$

 $\frac{{{\partial ^2}f}}{{\partial {x^2}}} = f(x + 1,y) + f(x - 1,y) - 2f(x,y) \text{，}$
 $\frac{{{\partial ^2}f}}{{\partial {y^2}}} = f(y + 1,x) + f(y - 1,x) - 2f(x,y) \text{，}$

 $K = \frac{{{\partial ^2}f}}{{\partial {x^2}}} + \frac{{{\partial ^2}f}}{{\partial {y^2}}} 。$

 图 5 船舶航行图像目标检测示意图 Fig. 5 Schematic diagram of ship navigation image target detection
3 结　语

 [1] 张杰. 基于灰色关联分析的舰艇大风浪中航行危险度评价[J]. 气象水文海洋仪器, 2013(1): 4. [2] 白鹤、张章、刘亚明、庞瑞、张振庚. 基于灰色关联分析的FDM工艺参数与制件精度关系研究[J]. 工程塑料应用, 2020, 48(11): 6. [3] 张杰. 基于灰色关联分析的舰艇大风浪中航行危险度评价[J]. 气象水文海洋仪器, 2013, 30(1): 17−20. [4] 张燕. 基于统计数学理论的舰船航行轨迹建模研究[J]. 舰船科学技术, 2020, 42(2): 53-55. [5] 李云成, 闻宏伟, 季正明. 一种基于灰色关联分析的空袭目标识别方法[J]. 地面防空武器, 2003(1): 37-38. [6] 张杰. 基于灰色关联分析的舰艇大风浪中航行危险度评价[J]. 气象水文海洋仪器, 2013, 30(1): 17-20.