【作者】 刘英东；

【导师】 卢光明；

【作者基本信息】 哈尔滨工业大学 ， 计算机科学与技术， 2008， 硕士

【摘要】 中医是中华民族的瑰宝。然而由于历史条件的制约,传统中医理论往往带有很强的主观性,其诊断手段也缺乏客观化诊断标准,故被许多人认为是一门经验科学,其科学性也屡遭质疑,严重阻碍了中医学的发展。如今,中医学的发展正强烈的呼唤着现代科学技术新方法、新手段的支持。如何在保持中医诊断自身优势基础上,通过现代技术手段改革与创新传统中医诊断学,实现中医诊断的客观化、标准化,将其纳入到现代科学技术发展的轨道是国人亟待解决的课题。电子鼻技术是当前计算机一个非常重要的研究方向。它的基本思想是利用模式识别方法对气敏传感器获取的气味数据进行分析。计算机技术的发展及电子鼻技术的成熟,使以电子鼻技术指导中医闻诊成为可能。运用电子鼻技术对传统闻诊信息进行研究,把闻诊研究建立在可靠的数学模型基础上,使其更准确、更客观地反映人体机能状态,从而提高中医理论学术水平和临床诊断能力,是闻诊研究领域发展提高的新课题。本项目在电子鼻原理的基础上,利用传感器技术、信号处理技术、模式识别技术,并结合中医闻诊的相关理论,开发出高精度、非侵入式、无伤害、快速准确的面向中医诊断的口腔气味检测系统(简称STCM),研究基于人体口腔气味检测与分析的中医诊断技术,以推动中医闻诊的客观化、标准化进程,具有巨大的社会意义和经济意义。为了满足闻诊诊断的要求,必须寻找能够检测口腔气味病理成分并具有足够敏感度的传感器。本文利用7个满足要求的气敏传感器组成传感器阵列;采用合适的信号调理电路和ADC、运放、CY7C68013单片机及USB模块等器件组成STCM硬件系统;用C51语言编写了单片机部分的控制软件;用Visual C++ 6.0编写了PC机部分STCM系统的应用软件,可实现实验的参数设置、实时显示、数据处理等功能。算法方面,分别在数据预处理、特征抽取和降维、BP神经网络识别方面进行了论述,并针对BP算法存在的不足,利用遗传算法对BP算法进行优化,从而建立针对病理气味数据分析的模型。更多还原

【Abstract】 Traditional Chinese Medical (TCM) is the huge treasure of the Chinese nation. But owing to the constraints of historical conditions, TCM theory strongly depends on subjectivity, and its diagnosis methods are short of objective criterion. Therefore it is thought to be empirical and its scientificalness is often questioned, which hinders the development of TCM seriously. Nowadays the development of TCM science is strongly calling for supports coming from new methods and new means of modern science and technology. It is a pressing problem to objectify and standardized TCM diagnosis and put it into the development of modern science and technology through innovating methods of modern technology and TCM diagnosis, maintaining the advantages of TCM diagnosis.At present, electronic nose technology (Enose) is one of the important researches in computer science and technology. Its basic idea is to use pattern recognition approaches to analyze the data gathered by gas sensors. The developments of computer and Enose technology provide a prerequisite for its steering breath diagnosis. When you do research on information of traditional smelling diagnosis by using Enose technology and mathematical module, it will reflect the state of physical function of human body and consequently improve both the academic standards theory and clinical diagnostic capacity of TCM. It is a new task in the field of developing and improving smelling diagnosis. This project, based on Enose theory, using sensor technology, signal processing technology, and pattern recognition technology, combining the correlated theory of TCM breath diagnosis, is aiming at developing high-precision, non-invasive, non-injury, fast and accurate breath odor detection system (STCM). It also plan to do research on TCM diagnosis technology based on breath odor detection and analysis, and accordingly promotes the objective and standardizing process TCM breath diagnosis. This also has great social and economic significance.Taking into account the special nature of smell, sensors, which are able to detect of oral pathology component and sensitive enough, are needed. An array of chemical sensors has been formed with seven gas sensors. Hardware of STCM system etc consists of preprocessing circuit, ADC, amplifier, CY7C68013 and USB etc. Software compiled with C51 in single chip, and software compiled with Visual C++ 6.0 in PC has the capability of experiment control, real-time display, and data processing, etc. In algorithm, areas of Data pre-processing, Feature extraction, BP neural network are discussed. BP algorithm is optimized by genetic algorithm, which greatly avoided local minimum of energy function by BP algorithm. At the final we established the analysis model of pathological odor data. .更多还原