基于水平集的弹创瞬时空腔X射线投影图像分割方法  

Level Set-based Segmentation of the Transient Cavity of Ballistic Wound in X-ray Image

作　　者：汪传忠[1] 史俊[2] 廖振强[1] 

机构地区：[1]南京理工大学机械工程学院,江苏南京210094 [2]南昌航空大学信息工程学院,江西南昌330063

出　　处：《测试技术学报》2011年第1期82-86,共5页Journal of Test and Measurement Technology

基　　金：国家973项目资金资助项目

摘　　要：弹创瞬时空腔多呈椭圆形,在近伤道入口和出口处,由于受到外界压力和进入气流的影响,瞬时空腔比伤道中央部小.根据目前弹创弹道学方面的知识和临床经验,往往不能准确检测出弹创瞬时空腔弱边缘.本文引入Chan-Vese水平集算法,利用自动阈值分割算法获取水平集初始轮廓线,利用零水平集曲线的平均边缘能量控制演化速度,利用零水平集停止条件自动获取最终轮廓线,解决了弹创瞬时空腔弱边缘难以自动检测的问题.The transient cavities created by gunshots are inevitably elliptic. The dimension of the cavity is smaller than the central wound channel, owing to the external and internal pressure nearing the entrance and exit of the wound channel. Since the detection of the weak contour is hardly accurate using the present ballistics and clinical experience, the application of the Chan-Vese level set method is introduced with the aid of automatic threshold segmentation to gain the initial contour, the average energy of the contour of the zero level set curve to control the evolution rate, the stop condition of the zero level set to automatically get the final contour.

关 键 词：弱边缘 图像分割 水平集算法 X射线投影图像 弹创瞬时空腔 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]