摘要
本文以CT图像为基础,针对肋骨的定位与追踪问题设计一个检测系统。系统可以自动检测出CT图像中的肋骨区域,并精确地勾画出肋骨的轮廓。首先对CT图像做预处理,消除噪声。然后采用OTSU方法自动选取合适的分割阈值,将CT图像二值化;采用形态滤波方法去掉二值化图像中的散点以及桥接裂缝,从而得到平滑轮廓的效果,将每个独立的连通区域标记出来;最后,通过提取出的特征对不同的肋骨进行分类,并且在不同的图像中追踪到相同的肋骨。
Based on CT images,this paper designs a detection system for locating and tracking ribs.The system can automatically detect the rib area in the CT image and accurately outline the ribs.First CT images are preprocessed to eliminate noise.Then use the OTSU method to automatically select the appropriate segmentation threshold to binarize the CT image;use the morphological filtering method to remove the scatter in the binarized image and bridge the cracks,so as to obtain a smooth contour effect,and mark each independent connected region Out;Finally,the different ribs are sorted by the extracted features and the same ribs are traced in different images.
作者
石志瑞
邓小炼
SHI Zhi-rui;DENG Xiao-lian(College of Science,China Three Gorges University,Yichang 443002,China)
出处
《南阳理工学院学报》
2018年第4期56-61,共6页
Journal of Nanyang Institute of Technology
关键词
CT图像
肋骨定位
肋骨追踪
分类决策
形态滤波
CT image
rib positioning
rib tracking
classification decision
morphological filtering