光学学报, 2017, 37 (10): 1011004, 网络出版: 2018-09-07   

含噪光学遥感图像海面弱小舰船目标检测 下载: 896次

Weak Ship Target Detection of Noisy Optical Remote Sensing Image on Sea Surface
作者单位
1 中国科学院长春光学精密机械与物理研究所小卫星技术国家地方联合工程研究中心, 吉林 长春 130033
2 中国科学院大学, 北京 100049
摘要
海面舰船目标的检测与识别对于海面监测与目标打击具有重要意义,弱小舰船目标由于缺少纹理信息且易受到海面阴影、噪声等因素的影响,使得目前常用的检测方法效果较差。基于通道分离与负值扩展对比敏感函数提出了的海面弱小舰船目标检测方法。该方法首先构建像素强度通道与噪声-边缘通道的多分辨图像尺度金字塔;之后,构建不同尺度空间下的负值扩展对比敏感度函数,调制对应各位置的权重;最后,利用各空间尺度系数加权获得两通道视觉显著性图像,通过通道差分处理实现了含噪图像中弱小舰船目标的快速检测。实验结果表明:与其他5种算法相比较,提出的方法具有较高的检测准确率(97.30%)、召回率(84.71%)及综合评价指标(94.49%),同时具备较强的抗噪声能力,适用于含噪海面光学遥感图像中弱小舰船的检测。
Abstract
The detection and identification of the ship target on the sea surface is significant for the monitoring and the target strike. The weak ship target with less texture information is affected by the sea surface shadow and noise interference, which results in the poor detection results with common algorithms. The weak ship target detection algorithm on sea surface is proposed based on the channel separation and minus extend contrast sensitivity function. Firstly, the multi-resolution image scale pyramid of the pixel intensity channel and the noise-edge channel is constructed. Secondly, the minus extend contrast sensitivity function under different scale space is established, and the weight of each position is modulated. Finally, based on the weighted spatial scale coefficient, the two channels' saliency maps are obtained. The weak ship target detection of noisy image is realized by channel differential processing. The experimental results show that compared with the other five algorithms, the proposed algorithm has high detection precision (97.30%), recall (84.71%) and F-Score (94.49%), and has strong anti-noise ability which is suitable to the weak ship target detection of noisy optical remote sensing image on sea surface.

宋明珠, 曲宏松, 金光. 含噪光学遥感图像海面弱小舰船目标检测[J]. 光学学报, 2017, 37(10): 1011004. Mingzhu Song, Hongsong Qu, Guang Jin. Weak Ship Target Detection of Noisy Optical Remote Sensing Image on Sea Surface[J]. Acta Optica Sinica, 2017, 37(10): 1011004.

本文已被 9 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!