摘要
实时目标跟踪是当前红外成像导引头图像处理中的关键技术，但是在目标跟踪的末端，匹配点漂移和滑动将直接影响目标跟踪精度，甚至使导弹丢失目标。本文深入分析了跟踪点滑动的原因，提出了基于特征点邻域的边缘模板匹配目标跟踪算法。该算法从红外图像中获取特征点，以特征点为中心选取参考模板，然后利用边缘检测算法获得边缘点集，最后使用改进的Hausdorff距离测度进行边缘点集的匹配，从而实现实时目标跟踪。仿真实验测试证明，该方法能够很好地解决跟踪点滑动和漂移问题，并满足实时性要求，跟踪稳定。
Real-time target tracking has become the key technology in image processing for an IR imaging seeker, but in the end of the tracking target, the sliding of tracking point and the locked frame excursion can make the missile miss the object. This paper explains in detail the research results on these questions. Then, a new real-time tracking algorithm for target is proposed based on the salient points of the edge template matching and tracking algorithm. In this algorithm, a small set of salient points are extracted from IR images using a Harris arithmetic operators. Therefore, the reference template can be selected according to these points, and then the edge points can be obtained by using the edge detection arithmetic operators. By measuring the improved Hausdorff distance between the salient points template and IR image, the object can be tracked in real-time. Experimental and simulation results show that this tracking algorithm can resolve two problems of existing algorithm very well： （1） the sliding of tracking point, （2） the locked frame excursion. Consequently, the algorithm could satisfy the requirement of real-time and tracking stability.
出处
《飞行器测控学报》
2008年第1期57-60,共4页
Journal of Spacecraft TT&C Technology
