面向智能视频监控系统运动目标检测的轮廓提取方法

A novel contour extraction method for motion object detection in surveillance video systems

针对传统的混合高斯模型方法易受干扰、运算量大的缺点,提出了一种应用于智能视频监控系统运动目标检测的轮廓提取方法.首先介绍了常用的运动目标检测方法;接着描述了传统的混合高斯模型方法,分析了该方法在目标检测方面存在的缺点,提出了一种新的轮廓提取方法.以过程为像素块混合高斯模型方法提取前景目标,采用数学形态学方法进行前景连通,Freeman链码寻找轮廓,Douglas-Peucker算法拟合轮廓,图像矩提取目标轮廓质心;最后对所提出的方法进行实验验证,并与传统混合高斯模型方法进行比较,实验结果证明,所提出的方法能更有效地滤除噪声,更准确地提取出目标轮廓.

Aiming at the problem that the traditional Gauss mixture model method has the shortcom ings of being interfered easily and needing a large amount of calculations, this paper presents a con tour extraction method used in intelligent video surveillance system for moving object detection. First this paper introduces the commonly used method of moving target detection; then it describes the tra ditional mixed Gauss model method, and analyzes the existed shortcomings in the target detection; it puts forward a new contour extraction method. This method uses pixel blockbased Gaussian mixture model approach to extract the foreground object, uses mathematical morphology method to achieve foreground connectivity, using Freeman chain code to find the outline, uses DouglasPeucker algo rithm to fit the contour, and uses image moment to extract the centroid of target contour. Finally, experiments verify the proposed method, and it is compared with the traditional Gaussian mixture model approach. The experimental results demonstrate that the proposed method can remove the noise effectively and extract the object contour accurately.