摘要
多媒体视觉运动轨迹标识在军事目标跟踪、机器人视觉、动态手势识别、交通信息监测等众多领域具有广泛的应用前景。针对当前方法由于受到多媒体视觉图像中复杂背景和光照、噪声等影响,无法提取获得更多有效的运动轨迹特征,导致运动轨迹标识结果与实际运动轨迹偏差较大,虚警率较高、提取耗时较长。提出一种基于人工蜂群算法的多媒体视觉图像运动轨迹标识方法,将采集获得的多媒体视觉监控图像中的运动目标进行二值化处理,并二值化处理后的运动目标图像进行背景自适应更新;采用自适应门限滤波对背景更新后的图像进行处理,滤除伪目标运动轨迹特,提取有效目标运动轨迹特征;采用峭度作为多媒体视觉图像运动轨迹特征独立成分分析的目标函数,利用人工蜂群算法对该目标函数进行优化求解,实现对目标运动轨迹的检测,并将目标运动轨迹标识出来。仿真结果表明,所提方法能够比较准确、快速地将多媒体视觉图像运动轨迹标识出来,且具有提取率高、正确率高、虚警率低的优点。
In this paper, we focus on a method to identify motion trajectory of multimedia visual image based on artificial bee colony algorithm. Firstly, the motion target in multimedia visual monitoring image was binarized and the background adaptive updating was performed on the motion target image after binarization. Then, the adaptive threshold filtering was used to process the image after background updating and the feature of pseudo - target motion trajectory was filtered. Moreover, the effective feature of target motion trajectory was extracted. Meanwhile, the peakedness was used as the objective function of independent component analysis of multimedia visual image motion trajectory feature. Finally, artificial bee colony algorithm was used to optimize and solve this objective function. Thus, the detection of target motion trajectory was realized and the target motion trajectory was identified. Simulation results show that the proposed method can identify the motion trajectory of multimedia visual image accurately and quickly. Meanwhile, it has the high extraction rate, high accuracy and low false alarm rate.
作者
肖瑜
XIAO Yu(Puyang Institute of Engineering, Henan University, Puyang Henan 457000, China)
出处
《计算机仿真》
北大核心
2018年第10期242-245,265,共5页
Computer Simulation
基金
关于我国寒冷地区外墙绿化节能技术基础研究(2016PYZYYY13)
关键词
多媒体
视觉图像
运动轨迹
标识
Multimedia
Visual image
Trajectory of motion
Identification