基于时变图像序列的脉搏信息提取

为获取全面客观的脉诊信息,本文提出一种全新的脉搏信息检测方法-脉搏触觉图像化法。本系统模拟中医脉诊实际,采用软性薄膜加压探头,由单一CCD摄像头采集得到脉搏动态时变图像序列。在算法实现上,从图像整体灰度变化入手,应用图像相关性和图像灰度重心变化分别提取脉搏波形图,并利用图像重心法分析得出不同人在不同取脉压力下的部分脉搏特征。研究结果表明,基于时变图像序列的方法能够有效提取脉搏波的幅度、形态和脉搏频率、节律等信息,为脉搏信息的检测和表征提供了新的手段和方法。

Information detection of pulse wave base on time-varying images

In order to obtain the whole and objective pulse information, a new method of pulse acquisition is presented-the imaging method of pulse tactile signals. A tactile detector was developed for simulating the practice of TCM pulse diagnosis, which has a soft film and exerts appropriate pressure on the wrist. The dynamic pulse images are collected by a monocular CCD camera from the wrist above the radial artery. Pulse waveforms are acquired by analyzing image correlation and image center. The pulse waveforms of different people under different pressures are obtained by detecting the center of image. The results show that the pulse information such as amplitude, shape, frequency and rhythm can be detected using time-varying image method, which provides a new approach to detect and express the pulse information.