基于变形模板技术的目标轮廓跟踪算法

为了适应目标旋转、尺度、场景光照等变化,利用B样条曲线表达目标轮廓,结合变形模板技术对运动目标的轮廓进行跟踪.在跟踪过程中,沿着模板曲线的法线方向检测目标轮廓,提高了检测效率.根据同一曲线上相邻点间的相关性,对检测所得的轮廓点集的坐标序列进行中值滤波,有效降低噪声干扰.将检测到的轮廓点集在形状空间匹配,使目标轮廓的形变限制在一定范围之内,有效抑制噪声和背景边缘特征的干扰.仿真试验表明,该算法能够有效得到目标轮廓,且具有较好的实时性.

Object Contour Tracking Algorithm Based on Deformable Templates

To adapting well to the rotation,scaling of the objects and variation of background illumination,in this paper we construct the contour Of object using B-spline curves,and track the moving object using deformable templates in image se- quence.When tracking,we detect the object contour along the template normals.And this method make the contour de- tection more fast.According to the correlation between the points nearby to each other and on the same curve,we use me- dian filter to process the coordinates of the contour points,in order to reduce the disturbance of noise.Restricting the con- tour shape in shape space by shape matching can effectively reduce the disturbance of noise and baekgrand edge.The re- sults of the simulation show that this algorithm can obtain the contours efficiently and fastly.