摘要
国土卫士视频监控平台被应用于大范围野外场景,因此存在干扰因素多、前景图像特征不明显等影响动态变化目标提取的不利因素。为解决这一问题,比较常见的方法是分析不同时间多张相片变化,但此法难以在图像中将噪声与运动小目标分离。据此,研发了顾及场景结构的运动目标提取方法。它有效去除了野外场景中树叶抖动等干扰因素影响,实现了监控平台下运动目标快速检测目的,提高了监控区动态变化效率。
The scene of application of video surveillance platform is a large range of outdoor scenes with many adverse factors such as interference,unobvious foreground image feature and others,which influence dynamics target extraction.The common method to solve this problem is to analyze the changes of multiple photos at different times,but it is difficult to separate the noise from moving small target in image.A new improving method called moving object extraction considered by scene structure,could get rid of interference factors of leaves shake in the wild scene,accomlict rapid detection of moving targets under monitoring platform,and improve the efficiency of dynamic change monitoring area.
出处
《现代测绘》
2016年第4期51-52,共2页
Modern Surveying and Mapping
关键词
国土卫士
视频监控
三维重建
运动目标提取
homeland guard
video surveillance
three dimensional reconstruction
moving object extraction
