﻿ 基于PQFT模型的遥感图像舰船目标检测方法
 舰船科学技术  2022, Vol. 44 Issue (15): 169-172    DOI: 10.3404/j.issn.1672-7649.2022.15.036 PDF

Research on ship target detection method in remote sensing image based on pqft model
ZHANG Li-jie, YIN Li-yuan, JIA Yi-xin
School of Computer Science and Engineering, Changchun University of Technology, Changchun 130012, China
Abstract: Ship target detection based on remote sensing images plays an important role in ship monitoring, maritime traffic management and military fields. In recent years, the research on fast and accurate recognition of ship remote sensing images has become a hot spot. Because there are a large number of interference signals in ship remote sensing images, this paper adopts a remote sensing image processing technology based on visual saliency PQFT model, combined with image particle filter algorithm, A new ship target detection method in remote sensing image is designed. A large number of image analysis data show that the algorithm has high accuracy and efficiency.
Key words: PQFT model     visual significance     remote sensing image     particle filter algorithm
0 引　言

1 视觉显著性检测算法PQFT的研究现状

 图 1 视觉显著性检测算法PQFT的原理图 Fig. 1 Schematic of visual saliency detection algorithm PQFT

 $\begin{split} & R(t) = r(t) - \frac{{g(t) + b(t)}}{2} \;，\\ & G(t) = g(t) - \frac{{r(t) + b(t)}}{2} \;，\\ & B(t) = b(t) - \frac{{g(t) + r(t)}}{2} \;，\\ & Y(t) = \frac{{r(t) + g(t)}}{2} - \frac{{|r(t) - g(t)|}}{2} - b(t) \;。\end{split} \text{}$

 $\begin{split} & RG(t) = R(t) - G(t) \;，\\ & BY(t) = B(t) - Y(t) \;。\end{split} \text{}$

 $\begin{split} & I(t) = r(t) + \frac{{g(t) + b(t)}}{3} \;，\\ & M(t) = \left| {\frac{{r(t) + g(t)}}{2} - I(t)} \right| \;。\end{split}$

 $\begin{split} & Q[u,v] = {F_1}[u,v] + {F_2}[u,v]{\mu _2} \;，\\ & F(u,v) = \frac{1}{{\sqrt {MN} }}\sum\limits_{m = 0}^{M - 1} {\sum\limits_{n = 0}^{N - 1} {{e^{ - \mu _{}^{2\pi (mv/M + u/N)}}}} } f[n,m] \;。\end{split} \text{}$

 ${f_t}(n,m) = \frac{1}{{\sqrt {MN} }}\sum\limits_{u = 0}^{M - 1} {\sum\limits_{v = 0}^{N - 1} {{e^{ - \mu _{}^{2\pi (mv/M + u/N)}}}} } {F_i}[u,v] 。$

 $S(t)=g\ast \Vert Q(t){\Vert }^{2} 。$

2 基于PQFT模型的遥感图像船舶目标检测方法 2.1 遥感图像舰船目标检测流程

 图 2 基于遥感图像的舰船目标检测流程图 Fig. 2 Ship object detection flowchart based on remote sensing image

1)遥感图像的预处理和海陆分离

 图 3 遥感图像噪声信号在RGB空间中的示意图 Fig. 3 Schematic diagram of remote sensing image noise signal in RGB space

2)区域提取和分类识别

2.2 基于PQFT的舰船运动目标跟踪技术

 ${q_\nu } = c\sum\limits_{i = 1}^n k \left[ {{{\left( {\frac{{{x_t} - {y_0}}}{h}} \right)}^2}} \right]\delta \left[ {b\left( {{x_i}} \right) - u} \right] \text{。}$

 ${p_u}(y) = 0.64\sum\limits_{i = 1}^n k \left[ {{{\left\| {\frac{{{x_t} - y}}{h}} \right\|}^2}} \right]\delta \left[ {b\left( {{x_i}} \right)} \right] \text{。}$

 ${w_i} = \sum\limits_{x = 1}^n \delta \left[ {b\left( {{x_t}} \right) - u} \right]\sqrt {\frac{{{q_v}}}{{{p_u}\left( {{y_0}} \right)}}} \text{。}$

 图 4 基于PQFT的舰船运动目标跟踪技术流程图 Fig. 4 Flow chart of ship moving target tracking technology based on pqft
2.3 遥感图像舰船目标检测的粒子滤波算法

 ${X_t} = {f_t}_{ - 1}({X_{t - 1}}) + {V_{t - 1}} 。$

 ${W_t} = {h_t}({X_t}) + {z_t} 。$

 $\chi = \frac{1}{{{N^2}}}\sum\limits_{i = 1}^N {\left\{ {{{\int {\left( {h\left( {{x_i}} \right) - 1} \right)} }^2}p({x_i})d{x_i}} \right\}} \text{。}$

 $p\left( {{X_{t - 1}}\left| {{W_{t - 1}}} \right.} \right) = \int {p\left( {{x_t}\left| {{x_{t - 1}}} \right.} \right)} p\left( {{x_{t - 1}}\left| {{W_{t - 1}}} \right.} \right){\rm{d}}{x_t} 。$

 图 5 粒子群滤波器的工作流程 Fig. 5 Workflow of particle swarm filter

 图 6 某港口船舶遥感图像的识别示意图 Fig. 6 Recognition diagram of remote sensing image of a port ship
3 结　语

 [1] 卢佩, 田敏, 张锐敏. 基于IS模型的视觉显著性分析及算法改进[J]. 石河子大学学报:自然科学版, 2015, 33(3): 385-390. LU Pei, TIAN Min, ZHANG Rui-min. Visual saliency analysis and algorithm improvement based on is model[J]. Journal of Shihezi University:Natural Science Edition, 2015, 33(3): 385-390. [2] 杨族桥, 陈跃鹏, 张青. 综合视觉注意模型的显著性局部特征提取算法研究[J]. 计算机科学, 2013, 40(8): 289-292. YANG Zu-qiao, CHEN Yue-peng, ZHANG Qing. Research on salient local feature extraction algorithm of integrated visual attention model[J]. Computer Science, 2013, 40(8): 289-292. DOI:10.3969/j.issn.1002-137X.2013.08.062 [3] 吴祯, 潘晨, 殷海兵. 视觉感知正反馈的显著性检测[J]. 中国图象图形学报, 2017, 22(7): 946-956. WU Zhen, PAN Chen, YIN Hai-bing. Significance detection of positive feedback of visual perception[J]. Chinese Journal of Image and Graphics, 2017, 22(7): 946-956. [4] 马翠红, 王毅, 毛志强. 基于梯度响应图与视觉显著性的快速实时目标检测[J]. 工业控制计算机, 2019, 32(5): 85-87. MA Cui-hong, WANG Yi, MAO Zhi-qiang. Fast real-time target detection based on gradient response graph and visual saliency[J]. Industrial Control Computer, 2019, 32(5): 85-87. DOI:10.3969/j.issn.1001-182X.2019.05.035 [5] 樊祥锰, 尚振宏, 刘辉, 等. 基于视觉显著性的均值漂移跟踪算法[J]. 传感器与微系统, 2017, 36(6): 130-133+137. FAN Xiang-meng, SHANG Zhen-hong, LIU Hui, et al. Mean shift tracking algorithm based on visual saliency[J]. Sensors and Microsystems, 2017, 36(6): 130-133+137. DOI:10.13873/j.1000-9787(2017)06-0130-04