摘要
为了解决单特征在目标跟踪中无法准确描述目标的问题,提出了一种多特征融合的实时目标跟踪方法。该方法将角点特征、轮廓特征融入传统的Camshift算法中,结合原有的颜色特征对目标进行描述。解决了传统算法易受同色物体干扰,抗遮挡性能差等问题。实验结果表明,该方法能够实现对目标的实时跟踪,当目标遮挡的时间较短时能够很好地识别目标,具有较高的鲁棒性。
To solve the problem of single-feature that cann't represent goal accurately in tracking,a real- time target tracking multiple features fusion method is presented. The method takes the comer feature and profile feature into the traditional Camshift algorithm, describes the target in conjunction with the original color and solves the problem of the same colored objects' interference and low anti-occluding in traditional algorithm. The experimental results show that the method can realize the real-time tracking of the target, and it can highly identify target with high robustness when the target is blocked for a short time.
出处
《信息技术》
2015年第1期85-88,共4页
Information Technology