摘要： 为了准确、快速地检测和追踪运动目标，提出差分法与改进的活动轮廓(snake)模型相结合的目标追踪算法。建立当前帧与前一帧的差分图像，判断目标的大概运动范围。在运动范围周围设定初始化曲线，用改进的snake模型进行目标拟合。采用帧差法测定运动范围，避免了snake初始化曲线人为选定的缺点，改进的snake模型使拟合结果更加准确。实验结果验证了该算法的有效性。

关键词: 差分, 活动轮廓模型, 目标追踪

Abstract: For the moving object can be accurately, quickly detected and tracked, this paper proposes an algorithm of object tracking based on the difference and the improved snake model. Moving region of the object could be gotten according to the current image and the last image. Initialized curve is set around the moving region and edge is fitted by the improved snake model. The drawback of the initialized curve is avoided by the moving region ensured by the difference. The edge is fitted accurately and quickly by the improved snake. Experimental results show that the algorithm is effective.

Key words: difference, snake model, object tracking

中图分类号: 

• TP391.41