•  
  •  
 

Abstract

The objective of this study was to assess the feasibility of identifying minerals on Mars using remotely sensed data. In the process we also investigated the effect of noise of Aerosol and dust particles on the spectra of Mars minerals. The remotely sensed data was obtained through modeling and simulation and compared to the lab spectroscopy of the specific minerals in order to make an accurate identification. A linear model was developed using MATLAB Random Number Generator to obtain a simulated image. Part of the information we needed for the linear model was pure pixel information of Mars which was obtained from Mars Spirit images. Random noise was added to the image in order to simulate a real world image. In addition to the random noise, a mathematical model was developed to represent the noise caused by aerosols and dust particles in Mars' atmosphere. The simulation was tested to ensure that it satisfied the appropriate model testing. Our results showed that our linear model was appropriate, and was accepted at a confidence interval of about 95%. The simulated image was then corrected from noise through iterations. The overall accuracy of the corrected image showed an improvement in classification by 25%. The signatures of the spectra of the two images were obtained and compared to the lab spectroscopy of specific minerals. The degradation of noise showed improvement in the spectral analysis of Mars data. The spectral analysis showed the presence of iron oxide, calcium oxide, and magnesium oxide leading to the conclusion that the image simulation is reliable in mineral spectral identification.

Share

COinS