Study Of Data Fusion And Visual For Multsource Goe-Special Data

With the national economy developed continuously and quickly, many multi-metal mineral resources face to the serious missing situation, which is not only restricted the development of economy, but also endangered the safety of national resources. Our country has already paid much attention about it. Mining exploring is more and more difficult with the mining exploring for a half century. Mine investigate in lie concealed, deeper and jumping-off district will be an important orientation of mining exploring. Therefore, it is necessary to look for a new method for information extracting。This become one of an important way to research the mining exploring by remote sensing ,geology,geochemistry,physical geography and so on. Having the knowledge of multi-geoscience information, we keenly feel that geography model is faulty. Because the complexity of the phenomenon and process in geoscience, how to obtain knowledge from multi-geoscience data become a special attention in academia. Geo science visual is being integrated with spacial analysis, data mining and knowledge discovery, which serve to spacial decision. However, the geoscience informatization plays an important role. Globosity spacial data mining, also named knowledge discovery, means finding information that is connotative, unaware and latent from globosity spacial database, extracting interest of spacial models and characteristics and the general relation between space and non-space data and other data characteristics which include in database. It transforms much original data to valuable knowledge. Geoscience visual is a concept from the integration of science calculation and geoscience, which is about visual expression and analysis of geoscience data. So, knowledge discovery of processing to multi-variable data fusion will be a mine investigate emphases to study. As more and more information sources and styles, it is necessary to deeply study and analyze the information content that is complex and associated. In order to obtain more enough results than simplex fusion, the technique of data fusion is often used to deal with multi-variable data, which can effectively eliminate uncertain factors of data and improve the expression of object and environment and the veracity of result or explanation analysis. Geoscience spacial data put multi-relativity and multidimensional enrich. Multi-relativity of geoscience spacial data put that spacial data of different subjects specialty have a definite relative which is qualitative or quantitative, while multi-space-time scale decided this correlativity of spacial data of different subject specialty put different degree of incertitude and complexity. Multi-relativity is a base of multi-geoscience spacial data fusion and data mining. Multidimensional enrich of geoscience spacial data put that the enrich of multi observation variable in the specialty data of different object field, such as spectrum dimension variable of multi spectrum/spectral in remote sensing data,multi-element variable of geochemistry data,multidimensional characteristic of physical geography and so on. Multi-variable and multidimensional data formed earth observation data volume of space-time dimension of multi-specialty and multi-measure of multi-variable dimension. The correlation of multi-variable geoscience spacial data and the content of integrated information extract for principle,model,method are keystones and difficulty in field of globe system in recent. The thesis is picked from "The crusted curtain action mechanism evolvement and form mine rules research in DaXingAnLing"and "Information extract method research of remote controlled mine based on GIS". Problems and research contents Comprehensive application study in remote sensing and multi-geoscience information is revealing various ways and different scale from attribution and characteristic belong to the object. Remote sensing makes use of differences between electromagnetism radiation to recognize and distinguish object, which has its good facet but also has limitation. It is proved that work along with the geology the increment of difficulties, remote sensing must learn the information with geology, physical geography, geochemistry, etc and is closely combined the comprehensive application with multi-geoscience information. In that way, we can know essence and its mutual contact of object indeed, acquire the satisfied geology application result. Mineral information has something to do with many metals minerals present diversity, different scale, obvious or obscure text information in the remote sensing image. It is out of place to describe it only use simple qualitative model. There are many problems:1) because of complexity of the mineral condition, differences of earth's surface overlays and resolution of remote sensing data, it can't confirm whether it is a significative mineralization only based on the information extracted from remote sensing data. In other words, it is difficult to quantitatively confirm the geologyinformation associated with miner information from component of matter, especially in covered and half-covered area. While regional geochemistry technology mainly make use of the element abnormal as direct symbol to explore mine, but because of abnormal intensity and scale are restricted from many geology factors, it is result in some weak and low abnormal information are neglected, some is exaggerate unusually, only according to some indeterminations in mine exploring use geochemistry abnormality. 2) The pure remote sensing data also just extract the information of certain" depth" mostly, so current research mainly limits at parts of geology information in earth's surface flat manifestation, spatial characteristics, but lack of depth research. 3) How choose the comprehensive research of remote sensing image processing method and multi-variable information, thus carry out to convert the two dimension surfaces into three dimension stereoscopic expression method, which has any problems existed. The end target of multi-variable data fusion and visual data mining is to build up the homologous special information reorganization and extraction system, it provide the analysis, transformation ability and extraction multiply theme information for research the complicate and huge system. Geology and geology phenomenon are functions of three dimensions space, and there are always existent multiply answer in abnormal information in remote sensing, geology and geology physical. Therefore adopt the fit evaluation, estimate models, extracting theme information considering many faces seem to be as importance in many ways. The thesis researched the method of multi-variable data fusion and visual data mining according to the data of remote sensing,geology,geochemistry,physical geography and so on. The basic idea is based on the theory and technique method of information fusion,data mining and knowledge discover. We discussed the geologic information extraction for the data of remote sensing,physical geography,geochemistry and so on in the experimental area. The technique route is: ⑴Research the scheme of matching and quantity among the sampling units of multi-variable spacial data according handle characteristic of multi-variable data about different spacial measures. ⑵Made visual data mining by the mathematical methods of multianalysis,principal components analysis,morphomatics,Hough transfer and so on. ⑶Fuse the multi original spacial data using the associate principal model based visual production of data mining. ⑷lidate the result of the visual data mining and data fusion in the experimental area.Main results This research carries on data fusion and visual data mining, set the typical silver mineral bed at Xin ba er hu Inner Mongolia for example. Main results from this research are listed as follows: 1. Extracted special feature information from remote sensing data based on Hough transform and morphology; 2. Designed and realized three-dimensional visualization use basic special data geology entity and phenomena such as geology, physical geography according to IDL, MATLAB. 3.Solving the rectify between pixel of remote sensing and geochemistry data sample units and quantity relationship, set up the diverse space data fusion model according to the connection rule method; 4.Build up the inside relation that alteration of remote sensing and the geochemistry abnormality, made use of special information such as alteration, structure extract from spectra data of remote sensing and regional geochemistry abnormal to carry out data fusion, advanced data fusion between geochemistry and remote sensing, even the flow of visual mining. Innovation 1.Solving the rectify between pixel of remote sensing and geochemistry data sample unit and quantity relationship, set up the diverse space data fusion model according to the connection rule method; 2.Build up the inside relation that alteration of remote sensing and the geochemistry abnormality, made use of special information such as alteration, structure extract from spectra data of remote sensing and regional geochemistry abnormal to carry out data fusion, advanced data fusion between geochemistry data and remote sensing even the possibility method for visual mining.