基于整体变分模型的地球化学数据降噪研究

Study on Geochemical Data Denoising Based on Total Variation Model

地球化学数据处理中如何有效地消除噪声是能否正确利用地球化学信息进行矿产预测的关键.借鉴整体变分模型在图像去噪上的应用,本文将其用于地球化学数据降噪处理,通过仿真实验和西藏自治区洞噶普铜矿的1∶1万土壤地球化学数据处理具体应用,结合累积频率法进行元素异常分带序列划分,可以发现经过整体变分模型降噪后的数据更能体现元素的空间分布特征,达到还原真实信息的目的,突出地球化学中成矿元素的异常,从而提高圈定矿产资源远景区的精度.

It is the key problem in predicting mineral resources that how to effectively eliminate the noise during geochemical data processing.In this paper,the total variation model into the geochemical data denoising has been taken,as it is in the the application of image denoising.The simulation experiments and application in 1∶10 000 soil geochemical data of Donggapu copper ore,Tibet shows the data reduced noise by the total variation model can reflect the spatial distribution of elements better than not reduced noise,and achieve the true information.Then the data reduced noise by total variation model can improve the accuracy of metallogenic prediction.