多尺度形态重构的微弱弹道轨迹检测  

Ballistic Trajectory Detection Based on Multiscale Morphological Reconstruction

作　　者：周宁[1] 李在铭[1] 周蓉[1] 

机构地区：[1]电子科技大学通信与信息工程学院,成都610054

出　　处：《电子测量与仪器学报》2008年第2期26-30,共5页Journal of Electronic Measurement and Instrumentation

基　　金：国家863高技术计划(编号:2004AA823120);国家自然科学基金(编号:10376005)联合资助项目

摘　　要：本文提出了一种多尺度形态重构的微弱弹道轨迹检测方法。该方法先用形态学灰值重构算法对序列图像进行背景估计并作二值化处理,提取出序列图像中的可疑目标。然后根据炮弹的凝聚特性对二值图像进行多尺度形态二值重构,消除二值图像中的大量虚警目标,最后根据炮弹的运动特性进行多帧数据融合检测弹道轨迹。仿真实验表明该方法能稳定检测出序列图像中的微弱弹道轨迹。A method for dim ballistic trajectory detection based on muhiscale morphological reconstruction is presented. The morphological gray reconstruction algorithm is used for background estimation, and the image is binarized to extract suspicious targets in single frame firstly. Then according to the shell aggregation characteristics, multiseale morphological binary reconstruction could be applied to eliminate the false targets. Finally according to the shell dynamic properties, ballistic trajectory is identified by multi-frame data fusion. Simulation experiment results show that the muhiscale morphological reconstruction algorithm has excellent practical value for dim ballistic trajectory detection in image sequences.

关 键 词：多尺度形态重构 弹道轨迹 微弱目标检测 结构元 

分 类 号：TN911.73[电子电信—通信与信息系统]