• •    

基于ViBe算法及分数阶微分边缘检测的运动目标检测

吴成1,李晓华2,周激流3,4   

  1. 1. 四川大学电子信息学院
    2. 四川大学
    3. 四川大学 计算机(软件)学院,成都 610064
    4. 成都大学 模式识别与智能信息处理实验室,成都 610106;
  • 收稿日期:2016-11-07 修回日期:2016-12-07 发布日期:2016-12-07
  • 通讯作者: 吴成

Extraction of moving target based on ViBe algorithm and Fractional Differential Edge Detection

  • Received:2016-11-07 Revised:2016-12-07 Online:2016-12-07

摘要: 摘 要: 本文提出一种结合ViBe算法和分数阶微分边缘检测算子的视频运动目标提取方法。首先利用ViBe算法检测视频中的运动目标;其次运用分数阶微分边缘检测算子来提取视频帧的边缘轮廓;然后将检测到的边缘轮廓与运动目标相“与”,从而获取运动目标的准确外边界;最后通过数学形态学处理,得到完整的运动目标。实验结果表明,相比于Canny,Sobel等算子,分数阶微分算子与ViBe算法相结合提取轮廓,可更有效的去除运动目标的阴影。从而可为后续的目标分类、识别、跟踪等打下了更好的基础。

关键词: 关键词: 运动目标提取, ViBe, 分数阶微分边缘检测, 形态学处理

Abstract: Abstract: This paper presents a video moving object extraction method combined with ViBe algorithm an-d fractional differential edge detection operator. Firstly, we use the ViBe algorithm to detect the moving objects in the video. Secondly, we use the fractional differential edge detection operator to extract the edge contours of the video frames. Then, the detected edge contours are ANDed with the moving objects to obtain the accurate moving objects Boundary; finally, the complete moving targets are get through the mathematical morphology processing.Experimental results show that compared with Canny, Sobel and other operators, fractional differential operator and ViBe algorithm combined to extract the contour, can be more effective to remove the shadow of moving targets. Which can lay a better foundation for the subsequent classification, identification, tracking and so on.

Key words: Keywords: moving object detection, ViBe, fractional differential edge detection, morphological processing