Sweat Multi-componet Analysis Bases On Near Infrared Spectroscopy

KCl and CO（NH₂）₂ compositions at surface of human body have a direct relationship with the human physiological condition.Researches on skin secretion detection were rarely carried out in the past times,especially the non-invasive measurement based on near infrared spectroscopy.Thus,the analysis for skin secretion composition has no unified standard.In order to detect the KCl composition and CO（NH₂）₂ composition of skin secretion at the normal body temperature 36.5℃quickly and easily,near infrared spectroscopy is adopted to measure the solution.After detection and analyzing the solution concentration of KCl and CO（NH₂）₂ based on near infrared spectroscopy,least squares method is used to fitting the absorption peak.Then two neural networks methods are used to apart the spectroscopy of compositions and analysis the concentration of unknown multiple compositions solutions.The main study is as follows:1,Measure NIR spectroscopy of sweat compositions.Establish a standardized sweat composition by near infrared spectroscopy data of KCl,CO（NH₂）₂ and multi-component sample.2,Qualitative and quantitative analysis of single sweat component.After studying the change regulation between NIR spectroscopy of component in sweat ingredients and concentration and temperature,based on a least squares analysis,the result that almost no influences on the spectroscopy with smaller temperature from35.0℃to 42℃is proved,and the spectrum increased with the enhancement of concentration at low concentration range.The determination coefficient of least squares method（r²）can reach to 96.28%.3,Quantitative analysis for multi-component of sweat.Apart the spectroscopy of compositions and analysis the concentration of multiple compositions solutions of KCl,CO（NH₂）₂.Results show that the regularity of absorption spectrum for multi-component of sweat varies with t concentration.while almost no relation with temperature.The separation of components and the judgment of concentration all are based on neural networks,and the r² all above 0.983.A theoretical basis is provided to monitor the condition of health and surface flexible device based on analysis of sweat multi-component.