摘要
Snake算法(动态轮廓模型)在图像处理过程中有着广泛的应用.提出基于归一化互相关系数的阴影检测方法,利用由帧间差分法得到的目标边界,自动设置Snake初始位置,采用贪婪算法得到最终目标收敛轮廓.实验结果表明,该算法能够快速有效地检测出多运动目标.
Snake algorithm, （active contour model） is widely used in image processing. At first, an algorithm of moving shadows detection, based on the normalized mutual correlation coefficient is proposed. Then using the method of inter-frame difference the object＇s contour points are obtained. The initial position of Snake is set automatically. The converged contour of the final object is got by using greedy algorithm. Experimental results show that the algorithm can quickly and exactly detect the multiple moving objects and suppress the shadows.
出处
《沈阳理工大学学报》
CAS
2006年第5期31-34,共4页
Journal of Shenyang Ligong University
基金
国防预研项目
关键词
归一化互相关系数
阴影检测
动态轮廓模型
SNAKE算法
normalized mutual correlation coefficient
shadow detection
Active contour model
Snake algorithm
