一种新的低信噪比红外舰船目标自动检测方法  

New automatic detection method for IR ship target on the condition of low SNR

作　　者：许旻[1] 牛照东[1] 陈曾平[1] 

机构地区：[1]国防科技大学ATR实验室,湖南长沙410073

出　　处：《红外与激光工程》2007年第z2期158-162,共5页Infrared and Laser Engineering

基　　金："十一五"装备预先研究项目(51301050103)资助

摘　　要：针对海天背景下红外舰船弱小目标难以提取目标形状的问题,提出一种新的目标检测方法。首先进行图像预滤波,检测海天线以确定目标潜在区域;然后对目标潜在区进行背景抑制,噪声抑制并增强目标细节信息;最后用快速区域生长检测出完整目标。实测数据表明,该方法能获取完整的目标形状信息,且实时性好。For the problem that IR-ship-detection is difficult to obtain the shape of the target,a new method is proposed.First,the image is filtered and the possible target region can be determined by locating the sea-sky-line.Then the background and Gaussian noise are restrained,the detail of the target is enhanced in the possible target region.The target is detected by the fast region-growing method at last.The experiment result shows that the method can efficiently detect integrated shape of the target,and has excellent real-time capability.

关 键 词：自动目标检测 目标增强 区域生长 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]