基于不变矩和灰度匹配的汽车跟踪算法

An Automotive Tracking Algorithm Based on Invariant Moments Matching and Gray Feature Matching

人工智能技术的引入为智能车辆提供了汽车目标跟踪的功能。但是传统的匹配跟踪算法不能够很好地实现窗口尺寸自适应调节功能,从而导致目标偏离甚至跟丢的情况。为了克服这个缺点,提出了一种窗口尺寸自适应的汽车跟踪算法。这种算法结合了不变矩特征匹配与像素灰度匹配两种算法。在MATLAB平台上,对此种混合算法进行了仿真。结果表明,改进的算法能在提高汽车跟踪准确性与实时性的基础上,很好地适应跟踪窗口的尺寸变化。

The introduction of artificial intelligence provides the function of automotive tracking for intelligent vehicles. As to traditional tracking algorithms based on image matching, adaptive changing the size of the tracking window cannot be achieved ideally. In order to overcome this weakness, an automotive tracking algorithm, which can change the size of the tracking window adaptively, was introduced. This algorithm combined invariant moments matching technique and gray feature matching technique. This blended algorithm was simulated on the platform of MATLAB. The results indicate that this revised algorithm can not only feature fast, precise and reliable, but also change the tracking window adaptively.