C^3I系统中战场视频镜头检测与关键帧提取  

On Shot Detection and Key-Frames Extraction of Battlefield Video in C^3I System

作　　者：郭忠伟[1] 宋枫溪[1] 王积忠[1] 

机构地区：[1]炮兵学院,安徽合肥230031

出　　处：《火力与指挥控制》2008年第S2期182-183,186,共3页Fire Control & Command Control

基　　金：学院专项基金资助项目

摘　　要：在C3I系统中,通过大量的战场视频进行态势信息提取以及目标识别一个重要环节就是镜头检测和关键帧提取。针对战场视频中有较多的光照条件改变,以及战场视频处理算法要简单快速等特点,给出了一种基于颜色直方图差的镜头检测和关键帧的提取方法。基本思想是同一镜头内的帧在颜色特征上没有显著变化,当连续帧的颜色直方图差超过阈值时则认为镜头出现了转换,对于同一镜头内的候选关键帧,计算之间的距离并与阈值进行比较来确定是否为关键帧。在VC++6.0的环境下,经过编程实现验证,该方法有效可行。A important tache of extracting situation information and target recognition from much battlefield video is shot detection and key-frames extraction in C3I system.The battlefield video has much illumination change,and the battlefield video processing algorithm must be simple and quick,so a practicable,based-on color histogram method of shot detection and key-frames extraction is presented.The basic idea is that there is no distinct color character change among the frames in the same shot,and when the color histogram difference between the sequential frames is more than the threshold the new shot appears.And the key frames is determined through comparing the threshold with the distance between the candidate key frames in the same shot.This method is proved effective through VC++6.0 programming.

关 键 词：战场视频 镜头检测 关键帧 直方图 目标识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]