序列图像中的目标区域跟踪算法的设计

An Algorithm Design for Tracking Object Region in Image Sequence

在测量高炮在基于雷达信号对飞行目标的自动跟踪精度时,采用的是在火炮身管上安装摄像机,然后对所摄取的图像进行处理并最终测算出跟踪精度。此种情况下,摄像机和目标均处于运动状态,对所获得的序列图像,分析了其特点,首先对图像进行滤波降噪,然后对图像进行了分割并使用数学形态学算法修正了分割结果的形态,基于图像中区域的特征,最后设计了一个自动跟踪图像序列中目标区域的算法,给出了判断是否发生误跟踪的标准,并对跟踪错误的情况进行了处理。根据对实际图像序列中目标区域的跟踪结果验证,本算法是可行的。

When measuring tracking accuracy of artillery controlled by radar signal,a camera is mounted in the barrel and then an image sequence can be captured.The camera and the flying object moves at the same time.According to this character,images captured by the camera are filtered to reduce noise firstly.Then images are segmented and algorithm of mathematical morphology is used to correct shape of the object.Lastly,based on features of regions in an image,design an algorithm which can automatically track and identify every object region in a sequence of digital image.It also provides a function which can judge if there is any wrong identification.According to the experience,this algorithm is available.