彩色图像序列中运动目标外轮廓的自动提取

该文提出了一种新的多运动目标精确外轮廓自动提取算法。算法的主要创新点是：采用指数滤波计算彩色图像的梯度强度，改善了彩色图像梯度局部极值点的定位精度，所得到的梯度强度和运动边缘位置信息供随后的轮廓迭代中使用，成功地避免了动态轮廓的过收缩。此外，充分利用轮廓的方向信息，又有效地克服了动态轮廓迭代中的外扩展。从而，算法自动地提取出一个真正紧贴的目标外轮廓。分析和实验表明，该算法抗干扰能力强，对复杂背景及不重叠物体，可以精确定位并提取出各个运动目标的外轮廓。

Contour Extraction of Moving Objects in Color Image Sequences

A new algorithm to extract accurate contours of multiple moving objects is proposed in this paper. The main contributions of this paper include two parts: firstly, exponential filter is used to calculate the gradient magnitude of color image and the localization accuracy of local maximum of gradient in color image is increased, the gradient magnitude and the motion edges are used subsequently in contour iteration to prevent the contour from shrinking successfully; secondly, the directions of contour are used to prevent the active contour from expanding effectively. As a result, the real compact contours would be extracted automatically by using this algorithm. The experiments show the high efficiency of this algorithm for contours extraction of moving objects from color image sequence with complex background.