空间弱小目标图像检测方法研究  

Research on Detection Dim Objects of Space Surveillance Images

作　　者：李德明[1] 王学伟[1] 

机构地区：[1]海军航空工程学院控制工程系,山东烟台264001

出　　处：《红外技术》2008年第10期559-562,共4页Infrared Technology

基　　金：国家"863"项目(2006AA0627)

摘　　要：针对背景缓慢运动的空间弱小目标图像检测问题,提出了序列图像背景配准方法及基于目标特征的空域滤波方法。通过对动态背景序列图像的背景恒星配准及背景消除,得到了目标图像。通过基于目标特征的滤波,并用质心法计算出每帧目标代替点得到轨迹投影图。最后使用Hough直线检测算法和最小二乘法提取目标轨迹。实验选择了实拍的序列图像,检测结果验证了该方法的有效性。For detecting moving dim spatial targets in dynamic background space surveillance images, a method that combined image preprocessing and spatial filter based on target feature is proposed. This method includes several steps： First, image preprocessing, it includes background stars registration, background mask generation and background subtraction. Second, enhancing image by spatial filter based on target feature, detecting the maximum 8-connected regions, getting target＇s coordinate through centroid method. At last, trajectory extraction through linear Hough transform and least square method. The effect of the detection method is verified by the result of experiments.

关 键 词：轨迹提取 图像配准 空域滤波 HOUGH变换 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]