论文部分内容阅读
基于Visual Basic 6.0,用图像信息统计法,设计了提取三维荧光等高线沿发射波长Em方向和激发波长Ex方向的一对特征谱的程序,介绍了提取方案和关键程序段。研究表明,该对谱具有较好的特征性和可计算性,解决了模式识别中特征参数提取的问题。基于I_(Em)和I_(Ex)谱,将各样本均与1号样本比较,得到了样本之间的欧氏距离d_(Em)和d_(Ex),通过平面距离聚类法和谱系聚类法分别对参芪扶正注射液与黄芪和丹参注射液的区分进行了鉴别,鉴别结果正确。此研究将图像信息数值化,解决了三维荧光图数值表征的问题,将特征谱与现代的数值分析技术结合起来,可对物质进行分类和鉴别。
Based on Visual Basic 6.0, a program of extracting a pair of characteristic spectra of three-dimensional fluorescent contour along the direction of emission wavelength Em and the direction of excitation wavelength Ex was designed by using image information statistical method. The extraction scheme and key program segments were introduced. The research shows that the spectrum has good characteristics and computability, and solves the problem of feature parameter extraction in pattern recognition. Based on the I_ (Em) and I_ (Ex) spectra, the Euclidean distances d_ (Em) and d_ (Ex) between the samples were compared with those of the first sample. Class method respectively Shenqifuzheng injection and astragalus and Danshen injection was identified, the differential results correct. This study quantifies the image information and solves the problem of numerical characterization of three-dimensional fluorescence spectra. By combining the characteristic spectrum with modern numerical analysis techniques, the material can be classified and identified.