对复杂边缘检测的Snake改进算法  

Effective Boundary Detection Method for Complex Target Based on the Snake Model

作　　者：赵保军[1] 李栋[1] 

机构地区：[1]北京理工大学信息科学技术学院电子工程系,北京100081

出　　处：《北京理工大学学报》2004年第2期162-165,共4页Transactions of Beijing Institute of Technology

基　　金：国家"八六三"计划项目(2002AA813032)

摘　　要：针对Snake原始模型对包含凹陷形状目标或信噪比较低的图像边缘检测结果差的问题,通过对复杂形状目标及含噪声图像的特性研究,从能量函数和迭代策略两个方面对Snake原始模型进行改进,增加了新的控制能量,并对各能量因素的影响权值进行了讨论,同时对蛇点采用了动态分布方法,以适应不同的目标形状特性.仿真结果表明,改进后的Snake模型较原始模型大大减弱了对蛇点初始位置的依赖,并在一定程度上有效地克服了图像噪声对迭代算法的影响,提高了对复杂目标的边缘检测性能.When image objects have non-convex forms or when there exists lower signal-noise rate in the image, the traditional Snake model often cannot detect the object edge effectively. The article attempts to study the characters of image that have complex shape target or noise, so as to improve the traditional Snake in the energy function and operation strategy. It gives new control energy function and discusses the right values of these functions, meanwhile dynamically arranges the Snake points for different target forms. Simulation proves that the new Snake model decreases greatly the Snake's dependence on initial conditions and overcomes effectively the influence of noise. This method enhances the snake's ability of detecting the object edge.

关 键 词：SNAKE模型 图像分割 能量函数 边缘检测 

分 类 号：TN911.73[电子电信—通信与信息系统]