﻿ 基于蚁群算法和神经网络的船舶图像压缩方法
 舰船科学技术  2022, Vol. 44 Issue (8): 165-168    DOI: 10.3404/j.issn.1672-7649.2022.08.035 PDF

Ship image compression method based on ant colony algorithm and neural network
LI Xiao-yan, SU Na
Qingdao Huanghai University, Qingdao 266555, China
Abstract: This paper proposes a new ship image compression method based on the ant colony algorithm and the neural network. Analyze the wavelet change rate of the image, construct a search space on the original image by segmentation method, and use rene search to match the data, thereby reducing the average value of matching, improving the average speed of fractal and encoding data, and collecting ship image data. The texture information in different directions is compressed by fractal prediction. The experimental results show that compared with the traditional methods, the ant colony algorithm and neural network ship image compression method consumes shorter compression time, higher compression ratio and stronger comprehensive ability, and is more suitable for practical work.
Key words: ant colony algorithm     neural network     ship image     image compression     compression method
0 引　言

1 蚁群算法和神经网络的船舶图像编码

 ${P_{}} = {\left( {\frac{{{k_{}}}}{{{k_{}} + f}}} \right)^2}。$ (1)

 $g\left( {{N_i},{N_j}} \right) = \sum\limits_{{N_j} \in {arces} \left( {{N_i},r} \right)} {\left( {1 - t\left( {{N_i},{N_j}} \right)} \right)} 。$ (2)

2 基于蚁群算法和神经网络的船舶图像压缩

 图 1 蚁群算法和神经网络的船舶图像压缩流程 Fig. 1 The ship image compression process of ant colony algorithm and neural network

 图 2 神经网络 Fig. 2 Neural Network

 $d\left( {{{\bar x}_1},{{\bar x}_2}} \right) \leqslant \frac{1}{{1 - s}}d\left( {{\omega _1},{\omega _2}} \right) 。$ (3)

 $\eta _{ij}^\prime = \frac{{V\left( {{x_j}} \right)}}{{\max \left\{ {1,\left| {{x_j} - {x_i}} \right|} \right\}}} ，V\left( {{x_i}} \right) = \frac{{\displaystyle\sum\limits_{l \in N{E_j}} {\left| {{x_j} - {x_i}} \right|} }}{8}。$ (4)

 ${\tau _{ij}}\left( {{t^\prime }} \right) = \rho \cdot{\tau _{ij}}(t) + \sum\limits_{k = 1}^N \Delta \tau _{ij}^k。$ (5)

3 仿真实验

 图 3 分割值域 Fig. 3 Splitting the range of values

 图 4 压缩实验结果 Fig. 4 Compression experiment results
4 结　语

 [1] 南敬昌, 臧净, 高明明. 改进蚁群算法的BRBP神经网络功放逆向建模方法[J]. 激光与光电子学进展, 2020, 57(1): 190-197. [2] 魏焕新, 张宏国. 深度神经网络结合蚁群算法的躲避攻击多目标对抗方法[J]. 计算机应用与软件, 2020, 37(11): 292-298. DOI:10.3969/j.issn.1000-386x.2020.11.047 [3] 谢云, 李志中, 唐广花, 等. 卷积神经网络的舰船图像快速无损压缩算法[J]. 舰船科学技术, 2020, 42(10): 62-64. DOI:10.3404/j.issn.1672-7649.2020.10.013 [4] 莫小芹, 魏进国, 付鹏浩. 基于图像处理技术的玉器自动比对鉴证溯源方法[J]. 工程技术研究, 2021, 6(23): 194-196. DOI:10.3969/j.issn.1671-3818.2021.23.059 [5] 孙洁, 于孟晗. 轧钢加热炉钢坯温度神经网络预测模型[J]. 国外电子测量技术, 2021, 40(9): 24-28. [6] 张达明, 何小海, 任超, 等. 基于自适应下采样和超分重建的图像压缩框架[J]. 太赫兹科学与电子信息学报, 2020, 18(2): 298-305. DOI:10.11805/TKYDA2018376 [7] 邓浩然, 吕成文, 陈东来, 等. 蚁群算法在土壤质地高光谱预测建模中的应用[J]. 土壤通报, 2021, 52(5): 1063-1068. [8] 周建新, 黄剑雄. 基于改进蚁群算法的加热炉温度控制研究[J]. 机床与液压, 2020, 48(18): 157-162,195. DOI:10.3969/j.issn.1001-3881.2020.18.023 [9] 李志军, 杨楚皙, 刘丹, 等. 基于深度卷积神经网络的信息流增强图像压缩方法[J]. 吉林大学学报:工学版, 2020, 50(5): 1788-1795. [10] 张红梅, 严海兵, 张向利. 结合半波高斯量化与交替更新的神经网络压缩方法[J]. 计算机工程, 2021, 47(5): 80-87.