基于机器视觉的病人多状态监测系统设计

Patient Condition Monitoring Method Based on Machine Vision Research

针对当前智能病人状态检测系统是对病人的单一状态进行自主检测和控制,实现的功能较为单一;设计了一种基于机器视觉的病人多状态识别监测系统,系统将前100帧的图像均值看成原始背景;按照不同的病态采集背景图像进行及时的调整;对当前图像和背景图像进行差分分析,采集差分图像,并对其进行形态学操作;按照二值图中运动目标的特征,分析病人处于哪种状态,通过扩展相应的功能模块,及时有效地检测出等待治疗、进行治疗、停止治疗等病人多状态的发生;以某医院常规病人状态视频进行系统测试实验,结果表明,系统可以实现实时检测,能够有效检测病人的多种不同状态,系统优化程度明显。

In view of the current intelligent patient state detection system is more to the patient＇s condition detection and control independently,the corresponding function has certain defects,this article designed a patients condition monitoring based on machine vision,extended the corresponding function module,can be timely and effective to detect the waiting treatment,treatment,and stop treatment the patient condition.Before the first 100frames of image mean as the original background;Again according to the different pathological gathering background image to adjust in a timely manner;Then the current image and background image difference analysis,gathering difference image, and carries on the morphological operations;Finally according to the characteristic of moving targets in binary figure,analysis of patients in a kind of state.Video through patient status,to experiment,results show that the method can realize real-time detection,and have high detection accuracy.