基于视频序列的运动目标追踪算法

摘要

An algorithm for moving object tracking based on video sequence is introduced, which integrates Kalman filtering and Meanshift algorithm to realize fast-moving target tracking. Kalman filter predicts the possible po- sition in the next frame, while the Meanshift algorithm locates the precise position of the Target. Actual tests show that this algorithm effectively overcomes the problem of missing target when the tracking target is moving fast, and that target tracking efficiency is improved significantly.%介绍了一种对视频序列中运动目标追踪的实现算法，该算法在运动目标检测的基础上，融合了卡尔曼滤波和Meanshift算法实现对快速移动目标的追踪。卡尔曼滤波对下一帧目标可能出现的位置做出估计，Meanshift迭代算子在估计出的区域对目标精确定位。经实际验证其有效地克服了传统Meanshift算法对于快速移动物体追踪可能出现的丢失目标的问题，目标追踪效果明显提高。