Hyperspectralmineral Mapping Algorithm Based On Non-parametric Statistical And Software Develop

Hyperspectral remote sensing technology is the field of remote sensing researchat home and abroad, especially remote sensing geological community project focus onresearch and development of high-tech high-tech applications. Due to the limitedtradition of spectral resolution multi-spectral remote sensing technology based onvisual interpretation difficult to effectively distinguish feature rich and unique spectralinformation. Since the last century from the more than three decades, hyperspectralremote sensing technology applications and feature spectrum of geologicalprospecting select and extract a lot of research work, and achieved encouragingresults. Studies have shown that identification of spectral information usinghyperspectral remote sensing image feature extraction geological body and thedelineation of the target volume has become possible. The development ofhyperspectral remote sensing technology, the development of new methods andtechniques for geological prospecting opportunities and new means.In this paper, through previous characteristics and stability of parametric studieson the basis of the spectrum of the wild rock and mineral, summed up the spectralabsorption characteristics and spectral characteristics curve shape can be used as awild rock and mineral spectral stability characteristics. Which the spectral absorptioncharacteristics mainly for the absorption trough or absorption peak wavelengthposition. Therefore, mineral spectrum recognition research direction focusedperformance in these three areas of the absorption of mineral spectra characteristicparameters, the overall shape of the spectral similarity measure based on the spectraof the physical characteristics of spectral knowledge model, this paper chose tospectrum as a whole form as the entry point of the study. The by stage research and analysis of the similarity measure based on the overallmorphological characteristics of the spectrum, first proposed based on thenonparametric Spearman’s rank correlation mapping techniques and Kendall taumapping techniques, and the use of the USGS spectral library data, from whicharbitrary selected minerals (olivine, the montmorillonite class, kaolinite, jarositesulfate minerals class, the hypersthene class and galena class) in different viewingconditions or different conditions of use of many mineral spectral curve of42spectralcurve data, these42spectral curve data into two groups, the test samples andreference samples. Mature at this stage SCM the mineral mapping of Figuretechnology and the proposed Spearman’s rank correlation fill mapping technology andKendall tau mapping techniques to identify the test sample with the reference sample,respectively, statistics each mapping techniques are not correctly identify the samplesnumber, and calculate the error rate of each of the three methods, Spearman’s rankcorrelation and comparative analysis, the proposed mapping techniques and Kendalltau the mapping technology misjudgment rates with mature at this stage SCM mineralmapping Figure technology false positives for comparison empirical analysis, wefound that the proposed two mineral mapping technology based on non-parametricstatistics, the effectiveness of the mineral mapping technology does not lose at thisstage mature SCM similarity metrics on the overall form of the spectrum.Experiments show that based on nonparametric Spearman’s rank correlation mappingtechniques and Kendall tau feasibility of mapping techniques.Developed on the basis of the results of the experimental comparison thenonparametric statistics mineral recognition software, and the software DexingCopper Mine of hyperspectral remote sensing image processing. Will Dexing CopperMine remote sensing image mapping results with Jiangxi Copper Company revisionof Dexing Copper Mine overall deployment diagram mine distribution diagram isvery similar to the comparison results, further evidence of the paper, based onnon-parametric statistics Spearman’s rank correlation mapping techniques andKendall tau feasibility of mapping techniques and non-parametric statistics mineralrecognition software practicality..Main innovations:This stage is based on research and analysis of the similarity measure of themorphological characteristics of the spectrum as a whole, for the first time putforward based on non-parametric statistics, Spearman’s rank correlation mapping techniques and Kendall tau mapping techniques, and experimental comparisonanalysis proved that these two technologies The feasibility.Spearman’s rank correlation mapping techniques and Kendall tau feasibility ofmapping techniques based on research and development of non-parametric statisticsmineral recognition software, and experiments prove the usefulness of the software.At this stage, the field of remote sensing applications in mineral exploration,spectroradiometer data and hyperspectral image data based, to carry out a spectralmatching, spectral modeling and geological mapping technology focus and emphasisinformation extraction technology research. This article is in this context to carry outrelated research, this study proposes a non-parametric statistical mineral mappingtechniques for remote sensing prospecting technical support, an increase of newmeans, and to a certain extent in the laboratory environment inspection, remains to beexamined and have been promoted in other regions.