基于图像匹配的血管外渗漏检测  

Extravascular Extravasation Detection by Image Matching

作　　者：刘坤涛[1] 郭振华[2] 钟鼎苏 

机构地区：[1]深圳市龙岗中心医院,广东深圳518116 [2]清华大学深圳研究院,广东深圳518116 [3]深圳市信冠医疗器械有限公司,广东深圳518116

出　　处：《医疗卫生装备》2012年第11期23-25,共3页Chinese Medical Equipment Journal

摘　　要：目的:提出一种非接触式检测血管外渗漏的方法,用于研制采用自动推注系统注射药液以进行CT造影剂增强扫描时监测血管外药液渗漏并控制推注的装置,降低血管外渗漏对患者的伤害。方法:基于图像匹配的非接触式渗漏检测方法,即利用彩色相机实时获取注射图像,并根据彩色图像的差异性,判断是否发生渗漏。结果:该文提出的算法可以检测1.5 mL造影剂的渗漏,在Matlab 2008R中处理的时间短于0.5 s。结论:基于图像匹配的非接触式自动血管外渗漏检测方法可以快速、有效地发现渗漏,能够用于研制非接触式检测血管外渗漏装置,提高自动推注系统使用的安全性。ObjeCtive To put forward a non-contact extravascular extravasation detection method for the auto injection system of contrast-enhanced CT scan, so as to develop the injection controlling device to decrease the injury of extravascular extravasation. Results Non-contact extravasation detection method based on image matching was used to obtain real-time injection images by the color camera and then determine if the extravasation happened. Results The method could detect the extravasation of 1.5 mL contrast and had a processing time less than 0.5 s for Matlab 2008R.Conclusion Non-contact monitor method based on image matching can detect extravasation effectively and efficiently, which may be used to develop a non-contact extravasation detection device to enhance the safety of auto-injection system.

关 键 词：造影剂 自动推注系统 血管外渗漏 尺度不变特征交换 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP274[自动化与计算机技术—计算机科学与技术]