基于遥感图像的舰船目标检测及特征提取技术

Ship target detection and feature extraction technology research based on remote sensing image

首先运用OSTU算法对舰船目标图像的陆域和海域空间分离,进而提出一种基于云干扰背景的高斯分布混合统计模型进行虚假目标的过滤。在实现舰船目标(ROI)检测的基础上,又根据舰船图像的特征提取形态特征和尾迹特征等,精确实现对舰船目标鉴别。数据表明,通过以上步骤的技术处理,能够很好地实现对遥感图像的舰船目标进行虚假过滤和快速的舰船特征提取。

In this paper,firstly we propose the spatial separation of land and sea by using the OSTU algorithm for the land and sea space ship target image. Morever,we proposed a gaussian distribution hybrid statistical model for false target filter which based on cloud interference background. When achieving target detection based on the ship target image,we attempt the ship image feature extraction based on target characteristics and morphological characteristics for achieving precise target discriminatio. Finally,the data shows that the technical processing through the above steps,this method can filtering false targets of remote sensing images and quickly extracts ship features.