改进的信息散度法在蚀变矿物分类中的研究

Improved information divergence method and its application in classification of altered minerals

为了有效获取待测光谱类型,采用信息散度匹配法对其进行分类处理。根据不同矿物具有不同的光谱特征曲线,建立不同矿物光谱曲线的标准光谱库,再采用信息散度法对待测矿物光谱与标准库中的光谱进行匹配,进而得出一种标准光谱库下的信息散度匹配法。采用该方法对实际光谱曲线进行分类处理,结果表明提出的方法能够有效对光谱曲线进行匹配,判别待测光谱所蕴含矿物类型,所得出的蚀变矿物类型与岩矿鉴定结果具有较高的吻合度,有利于提取蚀变矿物信息。

In order to identify the types of the spectrum under measurement effectively,the information divergence method is applied to classification of the measured spectra. The standard spectral library of different mineral spectral curves is set up according to different spectral characteristic curves of different minerals. The information divergence method is used to match the curves of the altered spectrum under measurement and the spectrum in the standard spectral library,so a new spectral information divergence method based on standard spectrum library is obtained. The new method is used to conduct the classification processing of the actual spectral curves. The results show that the proposed method can match the spectral curves effectively and identify the mineral types implied in the spectra under measurement. The altered mineral types got by the proposed method are consistent with the results of actual mineral identification,which is beneficial to the extraction of altered mineral information.