基于改进Camshift算法的NAO机器人目标跟踪

NAO Robot Target Tracking Based on Improved Camshift Algorithm

摘要:

针对Camshift算法应用于NAO机器人目标跟踪过程中,当目标受到相似颜色背景干扰或被物体遮挡时跟踪失败的问题,提出一种基于ORB特征检测和Kalman滤波多算法结合的目标跟踪方法。首先检测目标ORB特征点初始化搜索窗口,然后利用Kalman滤波作为目标运动状态的预测机制,以预测的位置初始化Camshift算法。利用Bhattacharyya距离判断跟踪窗口的收敛性,若受到背景干扰,则利用ORB算法对当前帧中的Kalman预测区域和目标模型进行特征点匹配,重新检测目标在视频帧中的位置。根据Kalman滤波预测目标被物体遮挡后可能的位置来更新预测器参数。实验结果表明,改进的算法能够在相似颜色背景干扰和目标遮挡的复杂环境下,连续稳定地跟踪运动目标。

Abstract:

When Camshift algorithm is applied to target tracking of NAO robot, the tracking fails if the target is interfered by similar color background or blocked by objects. A target tracking method based on ORB feature detection and Kalman filter is proposed. Firstly, it is to detect the target ORB signature point and initialize the search window. Then Kalman filter is used as the prediction mechanism of target motion state to initialize Camshift algorithm with the predicted position. The Bhattacharyya distance is used to determine the convergence of the trace window. If background interference occurs, the ORB algorithm is used to match feature points between Kalman prediction area and target model in the current frame, and the position of the target in the video frame is redetected. According to Kalman filter, the possible position of the target after it is blocked by the object is updated. The experimental results show that the improved algorithm can continuously and stably track moving targets under the complex environment of similar color background interference and target occlusion.