基于血液光谱的物种检测技术

Species Detection Based on Blood Spectrum

  • 摘要: 提出一种基于血液光谱数据识别血液来源物种的光谱检测技术.设计制作了一套可以采集抗凝管血样光谱的检测设备,使用该设备采集了649个血样的光谱数据,建立了一个小型的血液光谱数据集.为了满足物种识别的需求,建立了一种基于数据驱动的建模方法.使用该方法对样本数据集进行交叉验证,可以达到95%的识别准确率,并发现该准确率随着样本数量增加而升高.

     

    Abstract: A spectrum detection technique based on blood spectrum data to identify blood source species was proposed.A device which can collect the spectrum of the blood sample in the anticoagulant tube was designed,and the spectrum data of 649 blood samples were collected by the equipment.In order to meet the needs of species identification,a data driven modeling method was established.Use this method to cross-validate the sample dataset,the recognition accuracy rate can reach 95%,and the accuracy was found to increase with the increasing of sample size.

     

/

返回文章
返回