基于卡尔曼滤波的动态目标跟踪

Dynamic target tracking based on Kalman filter

为了实现工业相机对动态目标的准确、实时跟踪,提出了基于卡尔曼滤波的算法。通过创建背景模型来估计出当前背景,进而得到前景区域,并对前景区域进行相关处理,最后通过计算补集得到更新后的背景。此方法能根据不同场景信息调整前景与背景阈值,减弱背景区域造成的噪声影响,实时地根据场景变化快速、自动更新背景,并对每一位置的像素进行背景估计。通过在VS2010平台上结合JAI软件开工具包(Software Development Kit,SDK)调用Halcon函数库实现了卡尔曼滤波动态跟踪,其中JAI SDK用于开发千兆网相机,几乎支持所有千兆网相机。实验结果表明,该算法能够实现对目标的实时动态跟踪,实时性强,准确度高。

In order to realize accurate and real-time dynamic target tracking from the industrial camera, this paper puts forward an algorithm based on Kalman filter. By creating a background model, it estimates the current background,then obtains the foreground, and takes a morpho- logical image processing on foreground. Finally, the background will be updated by calculating the complement of the foreground. This method can adjust the threshold of foreground and background according to different scenarios information and abate the noise impact of the background region, and it can rapidly and automatically update the background according to the real-time changed scene. The algorithm was realized by Hal- con functions combined with JAI SDK upon VS2010. The JAI SDK for the development of Gigabit Ethernet cameras, supports almost all Gigabit Ethernet cameras. Experiment showed that it can real-timely and accurately keep track of dynamic target.