首页 | 官方网站   微博 | 高级检索  
     

基于视觉显著性的高分辨率遥感影像舰船检测
引用本文:叶秋果,闸旋,李海滨.基于视觉显著性的高分辨率遥感影像舰船检测[J].海洋测绘,2018(4):48-52.
作者姓名:叶秋果  闸旋  李海滨
作者单位:海军海洋测绘研究所
摘    要:通过研究、分析现有基于遥感影像的舰船检测算法,提出了基于Itti视觉显著度模型改进的高分辨率遥感影像海面舰船检测方法,该方法利用海面舰船相对于海面背景信息视觉关注高的特点,参考视觉显著度模型计算原理,对高分辨率遥感影像中的颜色、亮度、方向、纹理等特征进行有效提取,然后利用归一化方法将不同特征融合,形成有利于舰船目标提取得视觉显著度图,最后对生成的视觉显著度图进行阈值分割和精细化处理获得最终的舰船检测结果。与传统舰船检测算法相比,本文算法能够有效检测复杂海面背景下的舰船目标,减少了单一特征变化对检测精度的影响,提高了检测准确率,具有较好的抗噪性和较强的鲁棒性。

关 键 词:舰船检测  高分辨率  遥感影像  视觉特征  视觉显著性

High Resolution Remote Sensing Image Ship Detection Based on Visual Significance
YE Qiuguo,ZHA Xuan,LI Haibin.High Resolution Remote Sensing Image Ship Detection Based on Visual Significance[J].Hydrographic Surveying and Charting,2018(4):48-52.
Authors:YE Qiuguo  ZHA Xuan  LI Haibin
Affiliation:Naval Institute of Hydroaphic Surveying and Charting,Tianjin 300061 ,China
Abstract:Based on the research and analysis of the existing ship detection algorithms based on remote sensing image,this paper proposes an improved high resolution remote sensing image sea surface ship detection method based on the Itti visual salience model,which makes use of the high visual attention of the sea surface ship relative to the sea surface background information,refers to the calculation principle of the visual salience model,extracts the color,intensity,direction and texture features in the high resolution remote sensing image effectively,and then uses the normalized method to get the different ship features of the final image segmentation and the visual features of the visual image.Compared with the traditional ship detection algorithm,this algorithm can effectively detect the ship target under complex sea background,reduce the influence of single feature change on detection accuracy,improve detection accuracy,and have better anti-noise and strong robustness.
Keywords:
本文献已被 CNKI 等数据库收录!
点击此处可从《海洋测绘》浏览原始摘要信息
点击此处可从《海洋测绘》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号