一种多点式声带振动速率与振幅提取方法  

A Method of Multi-points Information Extraction of Vibration Velocity and Amplitude of Vocal Fold

作　　者：钱晶[1,2] 聂余满[1] 黄学勤[3] 雷建和[2] QIAN Jing NIE Yu-man HUANG Xue-qin LEI Jian-he(Institute of Intelligent Machines, Chinese Academy of Sciences, Hefei 230031, China College of Automation Engineering, Qingdao Technological University, Qingdao 266520, China Department of Otorhinolaryngology, the Third People＇ s Hospital of Hefei, Hefei 230022, China)

机构地区：[1]中国科学院合肥智能机械研究所,安徽合肥230031 [2]青岛理工大学自动化工程学院,山东青岛266520 [3]合肥市第三人民医院耳鼻咽喉科,安徽合肥230022

出　　处：《仪表技术》2017年第4期1-5,10,共6页Instrumentation Technology

基　　金：国家自然科学基金资助项目(61201400;61673369)

摘　　要：为在声带持续振动的情况下获取声带振动的平均速率与振幅,提出了一种基于图像特征识别的多点式信息提取方法。使用喉高速摄影设备采集图像序列后,通过确定目标振动区域和图像预处理的方法提高后续算法效率,利用声带振动特征获取声带轮廓点集,并根据向量积公式确定轮廓点集中有效点的位置,以此计算出声带振动的平均速率与振幅。实验结果表明,该方法实现了声带振动信息的量化,可为临床喉科医师诊断提供有效的参考信息。To acquire the average vibration velocity and amplitude of vocal fold information when vocal fold keeps vibrating,a method of information extraction based on image characteristics recognition is proposed.After the image sequences are extracted by throat high-speed photography,the efficiency of algorithm is improved by determining the vibration area and preprocessing images.The vocal fold contour points are collected by the feature of vocal fold vibration.Finally,the location of the effective point in the vocal fold contour points can be determined by the vector product;besides,the average velocity and the amplitude of vocal fold can be calculated.According to the results of the experiment,the quantification of vocal fold vibration information is realized by this method.It can also provide some effective reference information to clinical otolaryngology doctors.

关 键 词：声带振动 图像特征 速率提取 振幅提取 轮廓检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]