Based On FastICA And Variational Partial Differential Equation Of Geochemical Anomaly Extraction Method Research

In mineral exploration, geochemical data is an important source of information.one of the objectives of the geochemical data processing is to discard the dross and select the essential,discard the false and retain the true. ore-prospecting information extracted effectively from the geochemical exploration data,to reveal the geological nature, is a prerequisite for metallogenic prediction work. In this paper, introducing new geochemical anomaly extraction method and its application are key research issues.Independent component analysis (ICA) is a kind of new method in the field of signal analysis and processing in recent years. ICA can extract the useful information in the signal, working to provide more abundant information for subsequent analysis.ICA is an effective method of non-gaussian data analysis, this feature of Exploration elemental distribution of highly nonlinear pertinence, which is to be applied to the basic premise of the probe data processing. In this paper, ICA convergence performance of Fast independent component analysis (FastICA) algorithm geochemical data processing, in order to achieve reduction of original geochemical data element distribution of real purpose.In the process of geochemical data acquisition will produce sampling error, preprocessing error, error analysis and Superposition error Inevitably, to form the noise of the geochemical exploration data, eliminating the noise effectively has become a key to the correct use of geochemical information. Variational partial differential (PDE) method is another important means of data processing following the wavelet transform, it has been successfully used in the field of image de-noising. in this paper,based on nonlinear variational partial differential equations, exploring a new geochemical data noise reduction approach, attempts to use partial differential equations noise reduction in the total variation (TV) model of the classic noise reduction processing of geochemical data.Based on FastICA algorithm theory and variational PDE noise reduction ideas, processing the1:1million soil geochemical data of the Tibet Autonomous Region hole Karma ordinary copper, first extract the distribution of effective information prospecting geochemical data using the FastICA algorithm, according to the analysis of the results, the extracted information is in line with the actual geological study area, and then use the total variation algorithm to process the original geochemical data and geochemical data extracted by FastICA algorithm, achieved good noise reduction effect, From the comparative analysis of results, the noise reduction effect by the the FastICA algorithm extracted after geochemical data better, more in line with the actual situation of the mining area.