基于修正的Snake模型与特征对象法的多运动目标分割

Multi-movement Target Division Based on Correctional Snake Model and Characteristic Object Method

在分析常规的运动目标分割算法的基础上,以计算简单、速度快、能精确提取运动目标为原则,提出了基于特征对象的运动目标分割算法,并将该法与修正的Snake模型结合进行了运动目标外廓的精确提取.分析和实验表明:该算法需要调整的参数少,计算简单,速度快,抗干扰能力强,可以有效地消除多帧间运动目标的遮挡,在多运动目标不重叠的情况下,能精确定位多运动目标的外轮廓.

Based on the moving target division algorithm and the principle of simple calculation,great speed,and precise withdrawal of the moving target,the authors proposed an automatic division algorithm of moving target based on characteristics object,which combines the algorithm with correctional snake model to precisely withdraw the external outline of the moving target.Analysis and experiments have shown that the algorithm is simple and fast,few parameters need to be adjusted,and the capacity to resist disturbance is strong.The multi-frame moving object mask can be effectively eliminated.Without overlapping multiple moving objects,the external outline of each moving object can be precisely located.