Introducing the Iron Potential Zones Using Remote Sensing Studies in South of Qom Province, Iran ()
1. Introduction
The use of satellite images for mineral exploration has been very successful in pointing out the presence of minerals such as alunite, pyrophyllite, kaolinite, sericite, illite, muscovite, smectite, and carbonate which are important in the identification of hydrothermal alterations [1]. In addition, satellite remote sensing provides synoptic view, which is helpful in identification and delineation of various land forms, linear features and structural elements [2,3]. Hence, in primary stages, mapping geologic lineaments are important as long-range management for mineral exploration, because of their potentials for harboring ore bodies that are carried and deposited by ascending hydrothermal fluids [4]. Iran is located in the Alpine-Himalian orogenic and metalogenic belt and has high potentials for gold and copper and other base metal deposits. Satellite images have been used by many Iranian geologist researchers as the cheapest method for interpretation of the structural features and exploration purposes [5,6]. For instance, distinguished phyllic, propylitic and argillic alteration zones in the Urumieh-Dokhtar Volcanic Belt of Iran [7], separated a few hydrothermal alterations in the Valayesh in the northwestern Iran [6], distinguished phyllic and argillic alteration zones in the magmatic arc in southeast of Iran [8], separated hydroxyl and iron oxide minerals in the Yazd, central Iran [9], show that satellite images are useful for locating and interpretation of the structural features in the Zagros Structural Belt, in southwest of Iran [10]. The aims of the present study are to perform principal component analysis (PCA), apply spectral angle mapper (SAM), Least Square Fit (LS-Fit), Minimum Noise Fraction (MNF) and band ratio to map the iron zones on the ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) and ETM7+ satellite imagery data.
2. Geology Setting
Kermejegan area is located between longitudes 481493 - 487103 and latitudes 3787390 - 3792893 in the south of Qom Province, Central Iran (Figure 1). The studied area is a small central part of 1:100000 Kahak geological map. The litho logy of this part includes volcanics such as Andesite, Andesite Basalt to mega porphyry Andesite. Different types of Sandi tuff and Limestone in south and southeast parts and quaternary units are seen in East parts of the area. Hematitization and limonitization are seen in north and southeast part [11]. There is an iron potential zone in southeast of the area with Magnetite and Hematite mineralization in vein and vein lets type which is in detail exploration phase now, showing in the map with green point. Also there are 2 iron mine in southwest of the area which are shown with red points in the map. As the map showing, the main fault—Bidhend—is going throw the south iron potential and central part of the area (Figure 2).
Figure 1. The studied area in the map of Iran.
Figure 2. The geological map of studied area (1:100000 Kahak geology map).
3. Materials and Methods
3.1. ETM7+ Data
Principal Components Analisys (PCA) method was used for doing this phase.After processing with Geomathica software, RGB: (5 - 7) & PC2(5 - 7) & PC4(1,4,5,7) were used. The iron potential zones were shown in goldish yellow by this method (Figure 3).
3.2. ASTER Data
The ASTER is an advanced optical sensor comprised of 14 spectral channels ranging from the visible to thermal infrared region. It will provide scientific and also practical data regarding various field related to the study of the earth [12]. Various factors affect the signal measured at the sensor, such as drift of the sensor radiometric calibration, atmospheric and topographical effects. For accurate analysis, all of these corrections are necessary for remote sensing imagery. To this end, at the beginning of the path, data set in hierarchical data format (HDF) was used for this research and radiance correlation such as wavelength, dark subtract and log residual by ENVI4.4 software which is essential for multispectral images, were implemented. For ASTER image processing, MNF LS-Fit PCA SAM methods were used.
3.3. Integration of Data Layers in GIS
Data layers have to integrate in GIS for getting any results.For this reason the geological layer with the location of fault on that should be the main layer and then the iron oxide potential zones which recognized by remote sensing integrated with that. In next step iron mines and indexes integreted with them and at last the final map prepared.
4. Result and Discussion
4.1. Aster and ETM7+ Images Processing
4.1.1. Least Squares Fitting
Least Squares Fitting (LS-Fit) is the technique assumes that the bands used as input values are behaving as the variables of a linear expression and the “y” value of the equation, namely the predicted band information, gives us a calculated output value. This predicted band is what that band should be according to the linear equation. The minerals which are sensitive to a specific band are then differentiated from the features which are reflective to the other bands as well; just by taking the difference between the predicted values and the original values [13].
Distribution of iron oxide was created by using all the 3 visible and near-infrared (VNIR) bands as the input bands and VNIR-b1 as the modeled band (Figure 4).
4.1.2. Minimum Noise Fraction
The Minimum Noise Fraction (MNF) transformation is used to determine the inherent dimensionality of image data, segregate noise in the data, and reduce the compu-
Figure 3. 2 and 3 dimension processing image with check field of iron oxide zones.
Figure 4. ASTER image processing with LS-Fit method and check field.
tational requirements for subsequent processing [14-16]. MNF involves two steps: in first step which is also called noise whitening, principal components for noise covariance matrix are calculated. This step decorrelates and rescales the noise in the data. In second step principal components are derived from the noise whitened data. The data can then be divided into two parts. One part associated with large Eigen values and the other part with near unity Eigen values and noise dominated images. Using data with large Eigen values separates the noise from the data, and improves spectral results [15,16]. MNF analysis can identify the locations of spectral signature anomalies. This process is of interest to exploration geologist because spectral anomalies are often indicative of alterations due to hydrothermal mineralization [16]. MNF band 2, was used for iron oxide alterations (Figure 5).
4.1.3. Principal Component Analysis
The Principal Component Analysis (PCA) is a multivariate statistical technique that selects uncorrelated linear combinations (eigenvector loadings) of variables. Each successively extracted linear combination, or principal component (PC), has a smaller variance. PCA is widely used for alteration mapping in metallogenic provinces [17]. PCA technique has been applied in this study. An approach based on the examination of eigenvector loadings in each PC image is used for determining which image contains information related to the spectral signatures of specific target minerals. It is expected that the PC image that collects moderate to high eigenvector loadings for the diagnostic absorptive and reflective bands of the index mineral could be considered as the specific image for that mineral. If the loading of absorptive band is negative in sign, the target area will be enhanced in bright pixels, and if the loading of reflective band is negative, the area will be enhanced in dark pixels [18,19] (Figure 6).
Table 1 shows the eigenvector loadings for bands 1, 2, 3 and 4. Inverse of PC4 can show the areas with iron oxide.