Research On Some Key Technologies Of Three-dimensional Metallogenic Information Analysis

In recent years, the problem of the crisis mines is getting more and more serious in our country, such as the increasing demand for mineral resources which makes the deep and side part of concealed orebodies stereo quantitative prediction necessary. Currently, the multi-prospecting information of geological, geophysical, geochemical exploration and remote sensing, as well as spatial database and the GIS technology is used in mineral resources prediction. Based on vast amounts of multi-source data, synthesis and extraction of metallogenic information has become the key of concealed orebodies stereo quantitative prediction. Based on3D geological modeling technology, this paper carried out deep researches on3D spatial relation analysis of metallogenic information with the concealed orebodies quantitative prediction.For3D visualization and prediction of concealed orebodies, by using3D geological data modeling technology, for digital classifying and integrating the vast amounts of multi-source geoscience data of the crisis mines for many years, the integrated geological database can be established, and then3D geological models can be built.In order to quantitative extracting ore-controlling factors, the key technology of3D metallogenic information analysis is researched. For the complex shape of geological body, based on3D raster model, by using mathematical morphology and Euclidean distance transformation algorithm, the3D morphology of geological body can be analysed, such as extracting the surface morphology trend of geologic body, hierarchically extracting and quantificationally analysing the surface morphology undulance, and extracting the angle parameter between the geologic body surface and its trend surface. For the thinner shape of geological body such as geological interface, by building TIN model of it, the3D morphology of geological interface can be analysed, such as extracting the trend undulance of geological interface, the slope of geological interface, the angle parameter between interfaces and the distance field of geological interface.After extracting the ore-controlling factors mineralization distribution, the3D field models of ore-controlling factors are built by using the mathematical model of distance field. Based on the extraction of ore-controlling factors with nonlinear transformation of3D field model, the quantitative extraction of metallogenic information can be achieved.Taking the concealed orebodies quantitative prediction data of Dingjiashan lead-zinc deposit in Meixian town of Fujian province as an example to research the spatial relation analysis of metallogenic information, by using the traditional nonlinear transform to linear regression model and the BP neural network model, the nonlinear relationship between mineralization indicator and ore-controlling factors is respectively spatially analysed. And then the superiority and insufficiency of those models for metallogenic prediction are discussed by comparing two prediction results.