Research And Application On Geospatial Digital Spectrum Model

The research of geospatial digital features has become a hot and difficult problem in mathematical earth science research and digital geosciences at home and abroad.Its scientific essence is to measure internal structure of geological body and its relationship through spatial reconstruction and expression model,that is,from the perspective of mathematical abstraction,according to the relationship between number and shape to reveal spatial distribution and variability of geological objects,so as to excavate natural attributes and changing laws of geological object,easy to dig multi-target geoscience information.This paper makes a method research on digital features of geological bodies,which is a qualitative research on the basis of the previous research results.Compared with the traditional method,it focuses on the characteristics of hybrid geospatial big data fusion analysis.This method basically belongs to the nonlinear method,which makes spatial digital feature analysis and spatial variability measurement simultaneously.Based on the discussion of geological space definition,“four-properties” and "three-phases" structure division,this paper discusses in detail the problem of geological spatial distribution and variability,which provides a further theoretical basis for digital spectrum modeling.The selection topics of the paper belongs to the part of the research project of the topic “Uranium Resources Potential Evaluation Based on Big Data” in the National Basic Research Program(973 Program),the borehole logging data of the Daying uranium deposit in Ordos Basin and regional geochemical measurement data on the northern margin of the basin periphery are used as the data support of the application case.Then we make fusion analysis of geological big data to construct the comprehensive scale,and discuss the improvement and specific application problems of geospatial digital spectrum theory and methods,combined with regional geological information and previous research results.This paper constructs a spectrum model reflecting the distribution and variability laws of proximity uranium element combination on the basin periphery,and constructs a spectrum model reflecting the uranium mineralization anomaly in the basin,then analyzes and explains the results,which provides quantitative basis for gathering uranium capacity analysis of parent rock series bearing uranium based on geochemical observation data on the basin periphery and supplying uranium capacity analysis based on borehole logging data in the basin;finally,it provides a tentative basic research method for establishing evaluation system on sandstone type uranium resource potential.The main research results are as follows:1.Construct an optimization model for large samples and multivariate,and build the sample collectionIn general,there are a lot of invalid or false values in the geological original data,which are the main sources for forming geological noise,and will seriously affect the stability of the model.There are differences among the contribution of geoscience information due to sample size.Because of its large number and multilevel,the information richness for large samples is better than the small sample,but also because of the large number,it is bound to carry a lot of geological noise.Therefore,before the calculation of spatial spectrum,it must perform noise reduction for large samples and multivariate.Based on the previous research achievements,taking the principle of improvement,universality and priority into consideration,this paper constructs a nonlinear model on optimizing sample and variable.In the big data fusion analysis,it is necessary to construct a suitable large sample set,but not all original samples can enter the large sample set.We determine samples number,sampling interval,variable parameter selection,variable dimension,variable test quality and other indicators according to the research objectives,then construct the optimal sample space under the overall optimal conditions,and finally express a large sample data matrix.2.Provide a comprehensive scale calculation method,and achieve multi-factor combination analysisThe sample data is acquired from different disciplines and different observation test methods.Different data structure and dimension require us to make the preprocessing with dimensionless and uniformity scale so that all kinds of data are made "the same",conducive to fusion analysis.Then we simplify the multi-factor as a nonlinear independent factor which reflects the geological anomaly or mineralization anomaly,so that we can achieve the multi-factor combination analysis of geological anomaly and achieve the effect of spatial dimension reduction,while every independent factor gets specific geological meaning.On above basis,according to the structure of large sample data,we provide a comprehensive scale calculation method and calculate spatial comprehensive scale value of each sample,as the initial sample value of the large sample in the model calculation process.3.Construct a digital spectrum model(one-dimensional case)based on the borehole logging data of Daying uranium deposit in Ordos Basin,which provides the quantitative basis for determining the uranium anomaly spaceConstruct a geospatial digital spectrum model(one-dimensional case)based on borehole logging data of Daying uranium deposit in Ordos Basin,and the model is corrected and improved according to the previous research results and geologic background.The paper determines the mineralization anomaly scope of the geologic profile on four exploration lines through the comprehensive measurement for amplitude change of spectral density,autocorrelation trend distribution of variation function and original scale value.The results show that main performance for the characteristics of the mineralization anomaly are as follows:(1)high density,obvious density changes and high mean value of original scale;(2)large spatial variability fluctuates,obvious change and high mean value of original scale.We determine the mineralization anomaly scope according to the characteristics of these two groups,and fit it with the mineralization location of the known geological exploration profile.The effect shows that fit degree of mineralization anomaly area is relatively good,so it can be inferred that this determining method is feasible as a quantitative basis for determining mineralization anomaly.4.Construct a digital spectrum model(two-dimensional case)based on the regional geochemical measurement data on the northern margin of the basin periphery,which provides quantitative basis for spatial distribution and variability of proximity uranium element combination factorConstruct a geospatial digital spectrum model(two-dimensional case)based on regional geochemical measurement data,we make the spatial distribution and variation characteristics analysis of the independent uranium element combination factors to measure the geochemical spatial distribution and variability on Ordos Basin periphery.The results show that:(1)basin periphery exists rocks series providing uranium;(2)the uranium element combination factor belongs to the specific group mechanism for uranium element migration process,and each one is independent;(3)the spatial autocorrelation of each uranium element combination factor is imbalanced,which indicates that the spatial distribution of the factor migration process is imbalanced,and it proves indirectly that migration ability and providing uranium capacity of each uranium factor are different;(4)quantitative calculation of supergene uranium geochemistry can be used as a quantitative basis for determining providing uranium capacity analysis of parent rocks series bearing uranium on the basin periphery.