Abstarct: In this study, a method baxsed on multispectral satellite image processing and remote sensing has been proposed with the aim of segmenting the contaminated area and determining the level of contamination in the studied area. In the first step, the proposed method uses different bands of Sentinel 2 satellite data and creates a panchromatic image of the area of interest, and uses a combination of segmentation methods such as fuzzy pattern and superpixel to create a specific boundary baxsed on the brightness levels of all spectra in the satellite data. In the second step, optical images and radar data from Sentinel 1 and 2 satellites were used to re-examine the contaminated area, combining visible and radar spectra with the aim of reducing errors and increasing the accuracy of segmentation. In the final step, the contamination area identified in the previous two steps was examined using spectral analysis of the data. In this step, using Sentinel 2 satellite data, the similarity of the satellite signal reflection with data from different minerals was obtained for each area using the Pearson correlation coefficient. Finally, baxsed on the samples collected from the contaminated area, various mathematical relationships have been made to establish linear and nonlinear relationship between the reflectance of different channels and the percentage of contamination of each mineral. The innovations of this thesis include processing different spectra of the satellite optical image separately with the aim of integrating the final results and increasing the accuracy, using the combination of the gradient operator, fuzzy clustering and the normalized coefficient of variation feature in the first step. In the second part, separate processing of radar and optical data and integration of the results, simultaneous use of the curvlet transform and PCA and two efficient clustering methods to improve the segmentation of the final image, and in the third part, using the Pearson correlation coefficient to create a spectral databaxse of minerals with similar spectral patterns, providing an initial estimate of contaminant minerals in different areas using satellite images and proposing mathematical relationships, nonlinear and linear, to calculate the concentration of various pollutants baxsed on the average brightness intensity of satellite images in each area are the innovations of this thesis. The spectral pattern of the Pearson correlation coefficient was obtained baxsed on the similarity of more than 99% and according to the comparison with the practical results, the accuracy of the final relationships for estimating pollution was more than 96%. The RMSE error of the relationships in the worst case for the pyrite mineral was equal to 0.37.