一种改进的卡尔曼目标实时检测与跟踪算法

An Improved Kalman Object Real-Time Detection and Tracking Algorithm

对象跟踪已在视频监视、流量管理、视频索引、机器学习、人工智能等领域广泛使用,为快速、实时和准确对视频运动目标进行检测与跟踪,本文提出了卡尔曼滤波算法与粒子滤波器相结合的优化算法。该算法是在卡尔曼滤波算法框架下结合粒子滤波器,嵌入具有更高精度的运动目标跟踪方法用于估计和预测运动物体的位置,提升运动目标的检测与跟踪精度,并对改进算法和传统卡尔曼滤波算法进行了实验对比。实验结果表明,改进的卡尔曼滤波算法具有检测精度更高、实时性更强、跟踪效果更好的特点,对于视频图像中运动目标的检测与跟踪具有较高的应用价值。

Object tracking has been widely used in video surveillance,traffic management,video indexing,machine learning,artificial intelligence and other fields.In order to detect and track video moving targets quickly,real-time and accurately,an optimization algorithm combining Kalman filter algorithm and particle filter is proposed in this paper.Under the framework of Kalman filter algorithm,combined with particle filter,the algorithm embedded a moving target tracking method with higher accuracy to estimate and predict the position of moving objects and improve the detection and tracking accuracy of moving targets.The improved algorithm was compared with the traditional Kalman filter algorithm.The experimental results show that the improved Kalman filter algorithm has the characteristics of higher detection accuracy,stronger real-time performance and better tracking effect.It has high application value for the detection and tracking of moving objects in video images.