一种基于参数活动轮廓模型的高能闪光照相图像分割算法  

An Image Segmentation Algorithm for High-energy Radiography Based on the Parametric Active Contour(snake) Mode

作　　者：景越峰[1] 张小琳[1] 管永红[1] 刘瑞根[1] 

机构地区：[1]中国工程物理研究院,流体物理研究所,四川绵阳621900

出　　处：《光电子技术》2009年第2期92-97,共6页Optoelectronic Technology

基　　金：中国工程物理研究院科学技术发展基金资助课题(2007B01006)

摘　　要：针对高能闪光照相系统成像质量较差的特点,提出了一种基于参数活动轮廓模型(Snake模型)的闪光照相图像分割算法。该算法在传统高斯力Snake模型中引入包含图像区域信息的变力,以目标和背景两区域具有最小方差为准则,构建兼顾边缘和区域信息的外部能量函数。数值实验结果表明,该算法对初始轮廓位置不敏感,较好地解决了客体凹陷区域分割问题,能够实现对含噪声的弱边界闪光图像的自动分割。For it is difficult to deal with automatic segmentation of noisy weak edge high-energy radiography image with traditional data-driven segmentation method.An image segmentation algorithm for high-energy radiography based on the parametric active contour(snake) mode has been explored.This algorithm led into a variable force of two regions information incorporating foreground and background to the snake model.It is proved that this model is robust to initial contours placements and can segment noisy weak edge high-energy radiography image automatically.Besides,the testing for segmentation on concave boundary shows a significant effect.

关 键 词：高能闪光照相 最小方差 图像分割 参数活动轮廓模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]