Remote sensing methods cannot detect noble metals with a sufficient level of certainty, but widespread hydrothermal alteration minerals associated with mineralization can. A certain diagnostic group of hydrothermal minerals can be identified on the basis of their spectral signatures, with diagnostic features available mainly in the short-wave infrared (SWIR) part of the electromagnetic spectrum. Iron oxides and hydroxides generally have lower reflectivity in the visible region and higher reflectivity in the near-infrared, making the visible and near-infrared (VNIR) portion of the spectrum critical for the study of these minerals. Over the last 10-15 years there has been a significant technological breakthrough related to the improvement of quality and availability of space images. This made it possible to revise traditional approaches to mineral resources forecasting. The relevance of the task of large-scale forecasting and prospecting for minerals using remote sensing methods is conditioned by the necessity to obtain qualitatively new data on both new and previously studied objects without significant expenses for field works. Thus, the creation of new methods for predicting mineralization is an important task for prospecting geology. This work is based on the creation of a methodology aimed at mineral prospecting using remote sensing data (hereinafter referred to as VNIR). VNIR and SWIR ASTER spectral characteristics were used to identify geological features of the territory. The methods of spectral interpretation were studied: principal component analysis (PCA); spectral angle method (SAM); calculation of mineralogical indices (BR), which were tested on the territory of the Urals and Iran. The main purpose of the paper is to improve the methodology of mineral forecasting at the early stages of exploration. The paper describes the most highly effective methods of spectral analysis for the identification of hydrothermal alteration minerals. The selected methods are principal component analysis (PCA), spectral angle method (SAM) and calculation of mineralogical indices (BR). Methodology of analyzing high-resolution space images for solving forecasting and prospecting tasks in geology Initial space images were obtained from the USGS portal, preliminary processing including radiometric calibration, atmospheric correction was performed using ENVI software. After primary processing, a new data set is created based on VNIR and SWIR channels. Today multispectral data are optimal (in terms of price-quality ratio) for solving tasks related to both interpretation and determination of rock material composition. Moreover, well-proven methods of statistical processing and spectral analysis of images are applicable for processing of such data. The degree of absorption and scattering of sunlight of any object are directly related to its wavelength. The spectral image is the amount of interaction of solar radiation with the Earth's surface. So each mineral has its individual reflective characteristic (spectral curve), which is related to its chemical composition, degree and temperature of crystallization, and genesis. Spectral information can be obtained by laboratory measurement, by ground-based measurement with a portable spectrometer, and remotely from a spacecraft or airplane. Principal Component Analysis (PCA) is a robust statistical method that is used to suppress radiative effects dominant in all bands, resulting in improved spectral reflectance and enhanced representations of geologic features. The principal component method is designed to select uncorrelated combinations of features among correlated data, including in multispectral image processing tasks. The result of the tool is a multichannel image, where the number of channels is equal to a given number of components (one channel per axis or component in a new multidimensional space). The first principal component will be characterized by the largest variance, the second component will correspond to the second largest variance value not characterized by the first principal component, and so on. In most cases, the first three or four images of the resulting multichannel image produced by the Principal Component Method tool will describe more than 95% of the variance. The remaining individual raster channels can be discarded. Since the new multichannel image contains fewer channels and more than 95% of the variance of the original multichannel image remains intact, the computation will be faster while maintaining accuracy. PC1 depicts the largest variance since it is the largest cut that can be made through the ellipse. The direction of PC1 is the eigenvector and its magnitude is the eigenvalue. This method resulted in the eigenvectors of the covariance matrix (PC) presented in Table 1. 


Channel 1

Channel 2

Channel 3

Channel 4

Channel 5

Channel 6

Channel 7

Channel 8

Channel 9

PC 1










PC 2










PC 3










PC 4










PC 5










PC 6










PC 7










PC 8










PC 9










Figure 1. Laboratory spectra of muscovite, kaolinite, alunite, epidote, calcite and chlorite
PC5 is most likely to identify Al (OH) as it has the highest weights in channels 5, 6, 7. The sign shows the brightness of the pixels: a negative sign (minus) in mapping will show the desired values with dark pixels; a positive sign (plus) - respectively, with light pixels. As a result, the anomalies - zones of distribution of minerals of Al (OH) group were identified (Fig. 2).
Figure 2. Composite of PC5, PC6, PC4. The group of purple pixels characterize the presence of Al (OH) group. Northern Urals

Spectral angle method (SAM)

The method for comparing image spectra to individual spectra or to a spectral library (Boardman, unpublished data; CSES, 1992; Kruse et al., 1993a) is based on estimating spectral similarity in (n - dimensional) feature space by computing the spectral angle between them.

SAM computes the spectral angle between the closest set of spectra and the reference spectra, where A is the reference spectrum, B, C, D are the set of pixel spectra, and α is the spectral angle between the vectors.
Figure 3. Concept of the spectral angle method
This method, of course, has its advantages and disadvantages. The method is fast and easy to use, it allows comparing the space image spectra with laboratory spectra. The disadvantages of this method include insensitivity to brightness variations and probability of misclassification of similar spectra.

Having applied the SAM method to the territory of the Urals, we obtained a map of the Al (OH) mineral group (Fig. 4).
Figure 4. Muscovite and kaolinite obtained using the spectral angle method (SAM). Northern Urals
Using Spectral Library reference spectra as a comparison, the spectral angle method in the area revealed the possible location of muscovite and kaolinite and calcite.

Calculation of mineralogical indices

Mineralogical indices allow to identify the area of certain mineral groups. In order to compose mathematical operations with the channels of the spacecraft

spacecraft, the spectral characteristics of the mineral, namely the absorption zone, must be taken into account. The Al(OH) group was identified using the AL(OH)=(B4/(B6)*(B8*B8)) channel operation. ASTER data in channel 4 has a high reflectance spectral index, while channel 6 and 8 have absorption spectra.
Figure 5. Mineralogical index of the Al(OH) group. Northern Urals
Approbation of spectral interpretation methodology<br>
 Ural region. The Ural region has been studied to such an extent that no new large deposits are expected to be discovered, but small and possibly medium-sized deposits can be found, also largely due to the use of remote sensing methods.<br>
 exploration. In addition, a systematic reassessment of the resource potential of the Ural region is needed due to fundamental changes in the understanding of the geology of the area. Today, the Ural fold belt is viewed as a collisional structure with a complex paleogeodynamic history, rather than as a mobile geosynclinal zone, the doctrine of which was developed in the late 19th and mid-20th centuries. All major issues of the resource potential of the Urals have been described and solved from the geosynclinal theory. This applies to the issues of mineralogy, tectonics, and paleogeography of the area. Most of the material on geology requires systematic revision within the framework of new paleogeodynamic ideas about the formation of the Urals, in correlation with the history of development of paleotectonic movements, paleogeographic settings, ore genesis, etc., as well as with the history of paleotectonic movements. The performed studies with the use of innovative methods will allow to consider the history of formation of coal-bearing formations of the Ural region in a new way and assess the geological prospects of their productivity.<br>
 Administratively, the work sites are located in the Ural Federal District (Khanty-Mansiysk Autonomous District, Perm Krai, Sverdlovsk Oblast, Komi Republic). Geologically, the work area covers the Trans-Urals marginal trough and the West Siberian Plate.
 There are many known copper porphyry type deposits in the Central Iranian volcanic belt located in the northeast of the country (Fig. 6, p. 84). Iran has great potential for exploration of copper porphyry type deposits using remote sensing data due to its good physiographic conditions and high degree of outcrop. Thus, the region becomes an interesting area for studying the prospects of discovery of epithermal gold mineralization of copper-porphyry and vein type (Fig. 7, p. 84).&nbsp;<br>
 In the present study, two copper mining areas in the southeastern segment of the volcanic belt, Meiduk and Sar-Cheshme, were selected. ASTER space images were used exclusively.<br>
 This paper utilizes efficient spectral transformation techniques to identify hydrothermal alteration minerals. Methods of principal component analysis, calculation of mineralogical indices, minimum noise fraction (MNF) were used.<br>
 Principal component analysis is a method based on statistical approaches in which new coordinate axes (components) are created with directions corresponding to the directions of greatest scatter of the original data. Multispectral image two or three components are able to describe almost all variability of spectral characteristics. Thus, by canceling the other components, it is possible to reduce the amount of data without data loss.<br>
 Through principal component analysis, RGB composites PC5 (plant data), PC6 (Al(OH) group), and PC7 (Fe, Mg(OH) group) were obtained showing hydrothermal alteration in the Meiduk and Sar-Cheshme areas (Figure 8 on p. 84, Figure 9).<br>
 Mineralogical index (BR) calculations identified minerals such as muscovite. The index was calculated from the difference of channels 7/6, where bright pixels corresponded to the mineral Muscovite.<br>
 With Minimum Noise Fraction Transform, data reduction is achieved by eliminating channels containing the highest noise fraction. Minimum Noise Fraction Transform is a direct transformation. It also excludes channels containing connected (coherent) images (Boardman and Kruse, 1994).&nbsp;<br>
 The table of eigenvalues of all SWIR ASTER channels in Meiduk and Sar-Cheshme territory was calculated. As a result, composite RGB images were obtained, which are shown in Figures 13 and 14.<br>
 The hydrothermal changes in RGB images are similar to those obtained by PCA and BR. The results proved to be effective and consistent with the results of field and laboratory studies (Amin Beiranvnd Pour, 2011).
 Thus, it is confirmed that processing and analysis of ASTER space images is a valuable tool for mapping mineralogical units. The use of high-resolution space images allows us to interpret the distribution of mineral groups and hydrothermal alteration zones. Each of the methods described in the paper showed almost similar results in cardinally different regions, which proves the feasibility of their application.<br>
 Zones of mineral hydrothermal alteration associated with copper porphyry mineralization were studied using VNIR and SWIR channels. Spectral methods can adequately detect hydrothermal alteration zones using exclusively remote VNIR and SWIR data on a regional scale.<br>
 It is necessary to realize that complex use of both traditional and modern remote sensing methods is necessary to solve geological problems. However, a significant share of classical tasks related to the diagnosis of material composition, geological boundaries, elements of ore body occurrence can be solved with the help of remote sensing. Introduction of such innovative methods into the cycle of geologic and prognostic works will allow to significantly reduce time and financial costs.<br>
 Е.М. Шемякина - научный сотрудник геологического факультета МГУ им. Ломоносова, ассистент инженерной академии РУДН, ведущий специалист отдела ГИС - геолог ООО "Иннотер".<br>
 Список литературы:<br>
 1. Boardman, J. W., and Kruse, F. A., 1994. Автоматизированный спектральный анализ: геологический пример с использованием данных AVIRIS, север Грейпвайн Маунтинс, Невада: в материалах Десятой тематической конференции по геологическому дистанционному зондированию ERIM, Институт экологических исследований штата Мичиган, Анн-Арбор, MI, стр. I-407 - I-418.<br>
 2. Центр изучения Земли из космоса (CSES), 1992. Руководство пользователя SIPS, Система обработки спектральных изображений,<br>
 Версия 1.2, Центр изучения Земли из космоса, Боулдер, CO, стр. 88.<br>
 3. Hubner, H., 1969a. Геологическая карта Ирана Лист № 5, Южно-Центральный Иран: Тегеран, Национальная иранская нефтяная компания, масштаб 1:1000 000.<br>
 4. Hubner, H., 1969b. Геологическая карта Ирана лист № 6, юго-восточный Иран: Тегеран, Национальная иранская нефтяная компания, масштаб 1:1,000,000<br>
 5. Mars, J.C., Rowan, L.C., 2006. Региональное картирование филлитовых и аргиллитовых пород в магматической дуге Загрос, Иран, с использованием данных усовершенствованного космического радиометра тепловой эмиссии и отражения (ASTER) и алгоритмов логических операторов. Геосфера 2 (3), 161-186.<br>
 6. Амин Бейранвнд Пур, Мазлан Хашим, 2011. Идентификация минералов гидротермального изменения для разведки порфирового медного месторождения с использованием данных ASTER, Южный Иран, Journal of Asian Earth Sciences 42 (2011) 1309-1323.<br>
Fig.6 Mineralogical map