基于Kalman滤波器的运动目标跟踪算法

运动目标跟踪是计算机视觉中的一个典型问题,如何能准确快速的跟踪目标是研究的关键。提出了Kalman滤波器结合Camshift的改进算法。首先选取一段视频图像序列,通过背景差分法快速检测出运动目标,初始化搜索窗口,用Kalmam滤波器预测目标位置,再用Camshift迭代算法计算目标最优的位置,将结果作为Kalman滤波器进行下一次预测的估计值。实验表明,当目标被严重遮挡或受到同色背景干扰时,本算法仍能快速准确的跟踪运动目标。

Algorithm of Moving Object Tracking Based on Kalman Filter

Moving target tracking is always a typical problem in the field of computer vision. It has been involved in many areas of technology of video image processing,pattern recognition and artificial intelligence. So it has a strong re-search value. For the researchers, the key of the study is how to more accurately and quickly track the target. In this paper, the Camshift algorithm is improved by using Kalman filter. First of all, we should choose a video image se-quence,we can quickly detect moving targets by background subtraction,Initialize search window,and we need to pre-dict the target location with the Kalmam filter, then we can calculate the optimal target location with Camshift algo-rithm, finally, as a result of the estimated value of the Kalman filter for the next forecast. The experimental results show that when the target is blocked or interfere by the same color background,the improved algorithm is able to fast and accurately track the moving targets.