基于视频的人体运动目标跟踪研究

现实生活中的背景一般比较复杂,基于颜色特征的Camshift目标跟踪在复杂情况下实现效果并不理想。文中提出采用基于边缘直方图Camshift结合优化的Kalman滤波来预测视频中行人目标下一个可能存在的位置,并利用OpenCV相关函数,跟踪丢失的几率降低。先对RGB图像进行边缘检测计算边缘直方图,得到搜索窗口质心位置,再用优化的Kalman滤波对运动目标窗口进行预测更新,将信息反馈重新确定下一帧窗口质心位置,以克服遮挡及噪声的干扰。实验表明,该算法能较好地对人体目标进行检测跟踪。

Human Motion Target Tracking Based on Video

The background in the real life is relatively complex,and the effect of the Camshift target tracking based on the color feature are undesirable in complex situations.The modified Camshift based on spatial edge orientation histograms and optimized Kalman filters are adopted to predict the next possible position of the target in the video,and related OpenCV functions are used to reduce the failure of tracking.The position of search window is obtained by calculating the edge histogram and Kalman filter is used for forecast and feedback the information to determine next frame window position,which can overcome the occlusion and noise interference.Experiments show that the algorithm has good performance in the detection and tracking of the target.