纹理高阶分形特征的海面舰船目标检测方法  

Research on detection method of ship target on sea surface based on high-order fractal texture

作　　者：卢运娇[1] 罗莎[2] LU Yun-Jiao;LUO Sha(Mechanical and Electrical Engineering Department,Beihai Vocational College,Beihai 536000,China;Electronic and Information Engineering Department,Beihai Vocational College,Beihai 536000,China)

机构地区：[1]北海职业学院机电工程系,广西北海536000 [2]北海职业学院电子信息工程系,广西北海536000

出　　处：《舰船科学技术》2020年第4期37-39,共3页Ship Science and Technology

基　　金：广西高校中青年教师基础能力提升资助项目(2018KY1186);广西高校中青年教师基础能力提升资助项目(2019KY1593)

摘　　要：为了提高海面舰船目标检测效果,提出纹理高阶分形特征的海面舰船目标检测方法。首先分析海面舰船目标检测原理,并对海面舰船目标图像进行处理,然后提取海面舰船目标检测的纹理高阶分形特征,并引入卷积神经网络分析海面舰船目标的变化特点,从而建立海面舰船目标检测模型,最后通过仿真实验分析海面舰船目标检测的效果。结果表明,对复杂背景的海面舰船目标,本文方法不仅提升了海面舰船目标检测的准确性,解决了海面舰船目标漏检的难题,而且海面舰船目标检测速度明显加快,可以实现海面舰船目标实时监控。In order to improve the detection effect of the sea surface ship target,a method of detecting the sea surface ship target with high-order fractal texture is proposed.Firstly,the detection principle of sea surface ship target is analyzed,and the image of sea surface ship target is processed,then the high-order fractal feature of the texture of sea surface ship target detection is extracted,and the convolution neural network is introduced to analyze the change characteristics of sea surface ship target,so as to establish the detection model of sea surface ship target.Finally,the effect of sea surface ship target detection is analyzed through simulation experiments,and the results show that,For the sea surface ship target with complex background,this method improves the detection accuracy of the sea surface ship target,solves the problem of missing detection of the sea surface ship target,and the detection speed of the sea surface ship target is obviously accelerated,which can realize the real-time monitoring of the sea surface ship target.

关 键 词：纹理特征 分形理论 海面舰船 目标检测 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]