基于空间分布与统计特性的海面远景目标检测方法  

Marine farsighted target-detection method based on spatial distribution and statistical characteristics

作　　者：刘博文[1] 戴永寿[1] 金久才[2] 李立刚[1] 邵峰[1,2] LIU Bo-wen;DAI Yong-shou;JIN Jiu-cai;LI Li-gang;SHAO Feng(China University of Petroleum,Qingdao 266000,China;The First Institute of Oceanography,SOA,Qingdao 266000,China)

机构地区：[1]中国石油大学(华东),山东青岛266000 [2]国家海洋局第一海洋研究所,山东青岛266000

出　　处：《海洋科学》2018年第1期88-92,共5页Marine Sciences

基　　金：国家重点研发计划项目(2017YFC1405203);国家自然科学基金资助项目(61401111)~~

摘　　要：海面运动载体获取的可见光图像中,远景目标成像面积小、特征不显著、信噪比较低等特点增大了目标检测难度,现有方法难以同时取得较高的查全率与较低的虚警率。作者根据目标在图像中的空间分布与统计特性,提出了一种海面远景目标检测方法。首先结合直线特征与界线特征实现海天线检测,然后根据拉伊达准则实现明度图像中疑似目标像素与背景像素的分离,得到二值图像,最后对二值图像进行形态学处理,并根据疑似目标区域与海天线的位置关系,排除干扰区域。测试结果表明:所提方法具有较高查全率与较低虚警率,且实时性和鲁棒性较强。In visible-light maritime images, the area and characteristics of farsighted targets are not clear and the image SNRs are low. The existing target-detection method cannot obtain a high recall and low false alarm rate. Based on the target’s spatial distribution and statistical characteristics in the image, we propose a marine farsighted target-detection method. In this method, first, the sea–sky line is detected by combining the line and boundary features. Then, the Pauta criterion is used to obtain a binary image by separating the suspected target pixel from the background pixel. Finally, the binary image is morphologically processed and the interference area excluded based on the positional relationship between the suspected target area and the sea–sky line. Our experimental results indicate that the proposed method has a high recall rate and low false alarm rate, and is relatively robust and effective in real time.

关 键 词：海面远景目标检测 海天线检测 拉伊达准则 形态学处理 

分 类 号：P715[天文地球—海洋科学]