当红外图像中包含较强噪声时, C-V 模型水平集分割方法会产生大量冗余轮廓 同时, C-V 水平集采用偏

微分方程( PDE) 实现, 存在计算量大、分割速度慢的缺点。为此, 本文提出了改进的快速算法, 该算法保留了C-V

模型的全局优化特性, 并通过窗口滤波整合图像邻域空间信息来构建曲线进化的外部速度, 从而提高C-V 模型

的抗噪性并减少分割中产生的冗余轮廓 采用基于双链表的快速水平集算法来实现曲线的演化, 去除了传统算

法中的重新初始化和PDE 求解的过程, 减少了迭代步数, 提高了分割的速度-Infrared image contains strong noise the CV model level set segmentation method will produce a large number of redundant outline　CV level set partial differential equation (PDE), there are computationally intensive, split the shortcomings of slow. To this end, the paper proposes an improved fast algorithm, the algorithm retains the global optimization CV model and the integration of spatial information of image neighborhood filtered through the window to build the external speed curve evolution, thereby enhancing the noise immunity of the CV model and reduce redundant contour segmentation　evolution of the curve to achieve fast level set algorithm based on the double-linked list, in addition to re-initialize the traditional algorithms and PDE solving process, reduce the number of iterations, to improve the segmentation speed