Research On The Method Of Martian Finely Hydrous Minerals Detection With Hyperspectral Remote Sensing Data

Hydrous minerals are the good indicators of Martian geologic process and aqueous environment.The spatial distribution,abundance and formation mechanism of hydrous minerals should be better constrained to help analyze the Martian geologic and climatic process.The aqueous environment is commonly related to biological activity.Thus,the better understanding of hydrous minerals is also important for analyzing the habitable environment and finding the biological signatures.Furthermore,the detections of hydrous minerals with fine category can better constrain their formation mechanism,thus constrain the early geological and climatic conditions of Mars.Martian hydrous minerals have the characteristics of relatively low concentrations,scattered distribution,uncertain or unknown background minerals,together with the nonlinear mixture,instrumental noises and spectral variability,making it difficult to be identified and quantitatively retrieved.This thesis focus on solving above problems and difficulties,by studying the methods of mineral identification and quantitative retrieval with hyperspectral remote sensing data.The objective of this thesis is to obtain more accurate distribution and abundance of Martian hydrous minerals and derive their formation mechanism,providing important mineralogy information for analyzing geologic process.The major results and conclusions are as follows:1.The spectral features analysis of hydrous minerals.The mechanism of spectral characteristics of hydrous minerals was analyzed.The spectral features of minerals in shortwave infrared wavelength are mainly caused by the vibration of H₂O,OH^-,CO₃^2-or SO₄^2-.This thesis also discussed the spectral features of major hydrous minerals on Mars and introduced their parameterization.The Fe/Mg phyllosilicates are identified through absorption features at¹.4?m and¹.9?m,and².28-2.32?m.Al Phyllosilicates have strong absorption at².2?m.Polyhydrated sulfates display an absorption near 1.9?m,whereas monohydrated sulfates have an absorption near2.1?m.Both polyhydrated and monohydrated sulfates exhibit a feature at 2.4?m due to water combinations and/or overtones.Carbonates have the features at².3?m and².5?m with similar absorption strength.Hydrated silica exhibits wide absorption feature at².2?m.The spectral parameters are able to identify the hydrous minerals of broad families.However,the complex visual interpretation process is needed to determine the finely hydrous minerals.So,spectral parameters is time-consuming and subjective.2.Dynamic Aperture Factor Analysis and Target Transformation?DAFATT?method was proposed.At present,it is difficult to accurately identify the distributions of hydrous minerals with fine categories.For solving this problem,this thesis proposed DAFATT method to locate the hydrous minerals on Mars.First,using the Hyperspectral signal identification by minimum error?Hysime?algorithm to determine the important eigenvectors objectively.Second,a dynamic aperture were used to detect minerals from hyperspectral data on Mars.Finally,target library and modeled spectra were normalized to allow for RMSE comparison between minerals with different reflectance scales,allowing for robust assessment of fit quality.The pixels in which detections from all the differently shaped apertures intersect are considered as true detections.The validations with laboratory spectra data and previously well analyzed CRISM images indicates that DAFATT can accurately obtain the distribution of serpentine and magnesite and has the potential to be applied to other mineral identification on Mars.The initial application to CRISM images in Nili Fossae shows that there are magnesite and a lot of serpentine.It should be noted that the serpentine is first detected here.3.Hydrous Minerals abundance retrieval based on Sparse Unmixing?HMSU?method was proposed according to the distribution and mixture characteristics of hydrous minerals.First,the endmember library was constructed using laboratory spectra,In-scene endmembers and synthetic spectra to accommodate the spectral variability,spectral background,calibration artifacts,residual atmospheric effects and instrument noise.Second,each pixel of CRISM data and each spectrum in the endmember library were converted to single-scattering albedo and then normalized,making the mixture of mineral occur linear characteristic and correcting the different scale between CRISM data and laboratory endmembers caused by the different calibration,measurement condition etc.At last,the sparse unmixing algorithm was performed to find the optimal subsets from endmember library to best model the CRISM spectra and calculate the abundance simultaneously.The HMSU method has good performance based on the validations with laboratory mixtures,AVIRIS hyperspectral data on the earth and the previously well analyzed CRISM data.The initial application to CRISM images in Nili Fossae shows abundant serpentine with 95%confidence level,which is consistent with DAFATT results.4.The hydrous minerals mapping in the final three potential landing sites of NASA Mars2020 rover.This thesis combined the results of DAFATT and HMSU method to obtain the hydrous mineral abundance distribution in Jezero crater,NorthEast Syrtis and Columbia hills.A few new serpentine outcrops,which were validated with ratio spectra and comparison with previous studies,were detected.Near the alluvial fan in Jezero crater and the south of NE Syrtis,big magnesite plains were detected.A lot of scattered magnesite were detected in the NE Syrtis landing site ellipse.In the spatial scale of CRISM image,no hydrous minerals were detected in Columbia hills.5.Formation mechanism of hydrous minerals was determined.According to the results of Nili Fosse,Jezero crater and NorthEast Syrtis,there are assemblages of serpentine,magnesite and talc,which is in accordance with high concentration of magnesium olivine,indicating the presence of serpentinization and carbonization hydrothermal alteration.These regions may have habitable environment in the past and preserve the biological signatures.