欢迎访问《图学学报》 分享到:

图学学报

• 视觉与图像 • 上一篇    下一篇

基于灰度与纹理特征的空中目标识别与跟踪

  

  • 出版日期:2016-04-28 发布日期:2016-05-20

Research on Aerial Target Recognition and Tracking Based on Gray and#br# Textural Features

  • Online:2016-04-28 Published:2016-05-20

摘要: 视频监控技术在航天领域中得到广泛应用,已成为航天安全和管理的重要技术手段
之一。为了实现在环境影响及遮挡情况下对航天视频监测和管理中对空中运动目标的运动状态监
测,首先运用背景差分法与相邻帧差法相结合的方法检测出运动目标的变化区域,通过二值化、
形态学滤波等处理后得到运动目标,提出灰度特征与纹理特征相结合的方法识别运动目标。然后
采用改进的Mean-Shift 方法实现对空中运动目标的跟踪,完成对空中运动目标的安全监控。结果
表明:提取与选择的目标特征在识别中取得较满意的效果,针对空中运动目标的跟踪具有较好的
鲁棒性,该方法具有可行性。

关键词: 目标检测, 跟踪, 识别, Mean-Shift

Abstract: In recent years, the technology of video surveillance, widely used in the field of spaceflight,
has become an important technical means in spaceflight safety and management. This paper gives a
method to firstly detects change region of moving targets through background subtraction and temporal
difference. Detect moving targets through binarization and morphological filtering, thus moving targets
are identified combining with gray and textural features. Then Mean-Shift method to recognize and
track moving targets is adopted. Lastly an experiment video example to monitor is given. The result
shows that the feature of target has achieved good effect in the process of recognition. For the tracking
of spaceflight moving target the method has better robustness, and is feasible.

Key words: target detection, tracking, recognition, Mean-Shif t