32 Applications of Remote Sensing and GIS in Mineral Resources

Dr. Jitendra Kumar Pattanaik

 

CONTENTS

 

1.  Aim of the Module

 

2.  Introduction

 

3.  Distribution of ASTER and Landsat Channels

 

4.  Reflectance of different mineral in VNIR and SWIR

 

5.  Characteristics of Images

 

6.  Digital Image processing

 

7.  Case Study

 

8.  Conclusion

 

9.  References

 

 

1.  Aim of the Module

  • To learn different remote sensing images used for mineral resource study.
  • To learn digital image processing for mineral resource study
  • To learn how this techniques can be used for mineral exploration – a case study.

 

2.  Introduction

 

Industrialization and economic development of a country is largely depends upon the mineral resources of the country. For many countries mineral resources are the main source of national income. Hence mineral exploitation should be guided by the long-term national goal and perspectives. As we aware that rock (which is an aggregate of minerals) was first used by the human as tools for hunting and subsequently different minerals were used for wide range of applications. Expansion of human civilizations could not have been possible without mineral resources. Directly or indirectly human completely depend up on the mineral resources.

 

Minerals are generally classified into three categories: metallic, non-metallic and fuel (energy). Metallic minerals are again categories into two types: a) ferrous (containing iron), e.g. iron, manganese deposits etc.; b) non-ferrous e.g. bauxite, base metal such as copper, lead, zinc etc. Non-metallic deposits also called as industrial minerals e.g. salt, clay, sulfur, limestone etc. Fuel or energy minerals deposits encompass coal, petroleum, natural gas and minerals used for nuclear energy. Some of the minerals are very rare and has various applications; hence it kept under precious mineral categories e.g. gold, silver and platinum etc. Other than above categories, minerals or rock are also used as building stone or road materials. Based on the mineral exploration result mineral deposits of any country are classified into two types a) mineral reserve and b) mineral resource. Mineral deposit which can be economically and legally extracted or produced at the time of the reserve determination is called mineral reserve. Mineral resources are the mineral deposits which will be economical in near or indefinite future. Mineral resources are either measured or indicated.

 

As per the mineral use data (2017) from the USA government agency (National Mining Association, the U.S. Geological Survey and the U.S. Energy Information Administration) every person of America requires 3.188 million pounds of minerals, metals and fuels in their life time keeping average life expectancy of 78.8 years (www.mineraleducationcoalition.org). Details of mineral resources are listed in the Table 1. This number will differ for other country depending up on the industrialization, development and lifestyle of the citizens. Similar to the above estimation, USA requires 40,641 pounds (www.mineraleducationcoalition.org) of new minerals every year per person to meet his or her requirement. This number varies every year and estimated based on the previous year mining and production.

 

In India mining sector is an important segment of national economy. Miningactivity provides raw materials to many industries such as iron and steel, thermal power, cement, fertilizers, petro-chemicals, glass and ceramics, electrical and electronics, building materials, precious and semi-precious stone etc. India is rich in iron, copper, chromite, bauxite, coal, crude oil and gas, base metal, manganese deposits. For many years exploration of various mineral/ore deposits activities was conventional type with restricted input fromgeochemical, geophysical and remote sensing techniques.

 

But now trend has changed and multidisciplinary approach being applied for mineral exploration. Minerals are non-renewable resources; hence its restricted use, recycle and reuse have been adopted by many developed country. Day by day demand of mineral resources increases to meet human requirement. Therefore mineral mapping, exploration and management are very important for any nation. The distribution of mineral resources on the earth is highly inhomogeneous and its mapping requires rigorous geological field survey as well as need more time. Prior to the remote sensing and GIS techniques all the mineral or lithological mappings were carried out only by the field survey. Advancement in remote sensing and GIS techniques provided ample opportunities to do the detail mineralogical / lithological mapping of an area in lesser time and field work.

 

Table 1: Shows the requirements of minerals, metals and fuels per person for his or her lifetime (keeping life expectancy of 78.8 years) in USA. This number varies for different years, e.g. in 2006 – 1678 metric tons, 2008 – 1633 metric tons, 2011 – 1343 metric tons, and 2017 – 1446 metric tons. Numbers are rounded off to the nearest number. (Source: www.mineraleducationcoalition.org). Here amounts are given in kg/g/l/cubic meter units which are converted from lbs/oz/gallon/cubic ft as reported in the source.

 

 

Remote sensing and GIS techniques have been widely used in the various fields of geological sciences. Among all, preparation of mineral resource map and exploration are important application.Mineral exploration and lithological mapping is a time consuming, laborious and required extensive field work. But use of remote sensing technique has drastically reduced the field work. In the early days of remote sensing aerial photography was extensively used by the geologists for mineral exploration. The advancement of satellite imagery in the subsequent years provided better opportunity to use remote sensing techniques for mineral exploration even in the inaccessible areas. The lunch of first earth resource satellite (Landsat 1) in 1972 provided a new avenue for the geologist to interpret digital satellite image for mineral exploration. In the recent years using high resolution multispectral satellite and airborne digital image geologist are exploring elusive potential mineral deposits which are cover by the vegetation and Quaternary deposits.

 

Type of sensors used for mineral resource studies: Remote sensing data /information about an area or object were collected by passive or active sensor mounted in a satellite or aircraft.

 

a) Passive sensors collect data using reflected or transmitted parts of the electromagnetic (EM) spectrum (Figure 1), which rely on solar illumination of the ground or natural thermal radiation for their source of energy respectively.

 

Important sensors for mineral resource studies are:

 

1.  Landsat Multispectral Scanner (MSS);

2. Landsat Thematic Mapper (TM) utilizes additional wavelengths, and has superior spectral and spatial

resolution compared with MSS images; ETM – Enhanced Thematic mapper,

3.Advanced Space borne Thermal Emission and Reflection Radiometer (ASTER) on

4. SPOT, a French commercial satellite with stereoscopic capabilities

5. The Indian remote sensing (IRS) satellite series combine features from both Landsat MSS/TM sensors and the SPOT HRV sensor.

b) Active sensors use their own source of energy. They emit energy and measure the intensity of energy reflected by a target. Some examples are Radar (microwave) and Lasers (European Space Agency 2004).

 

 

Landsat Satellite series MSS/TM provides multispectral imageries having a wide applicability in geological studies. ASTER sensor provides higher spatial and radiometric resolution for similar part of the electromagnetic spectrum. Remote sensing data gathering system records part of the electromagnetic spectrum (Figure 1). Some part of the EM spectrum get absorbed or interrupted by the atmospheric gas (O3, CO2, H2O etc.), hence data is restricted to a particular band of the EM spectrum. Part of the ultra violet (UV) radiation gets absorbed by the O3 and different segment of reflected & Thermal IR wavelength areinterrupted by the H2O (water vapor) and CO2. Each remote sensing satellite systems are designed to record a particular band with different spatial and temporal resolution (Figure 2). Hence it decides the applicability of the data to different natural resource studies.These data can be used to extract information about the lithology or rock composition, land surface structure, shallow subsurface rock type etc. For mineral resource study the reflectance / emissivity of the mineral is used and information were extracted from the remote sensing data.When light incident the mineral or rock surface certain wavelength of light gets absorbed and some are reflected depending up on the chemistry and crystal structure. Absorption of energy is mainly due to electronic (presence of color center, charge transfer, conduction band and crystal field effect) and vibration process of molecule (Whateley, 2006).

Figure2.Some parts of the EM spectrum are wholly or partly absorbed by atmospheric gases. Atmospheric windows where transmission occurs are shown and the sensors of different satellite system that use these wavelengths are indicated. EM Spectrum showing the majority of the data-gathering wavelengths. (Modified after Whateley, 2006)

 

  1. Distribution of ASTER and Landsat Channels: ASTER has three groups of channels: a) three recording (1, 2 & 3)of visible and near infrared radiation (VNIR) at a spatial resolution of 15 m; b) six recording portions (4, 5-9) of shortwave infrared radiation (SWIR) at a spatial resolution of 30 m; and c) five recording (10-14) thermal infrared radiation (TIR) at a resolution of 90 m (figure 3).Whereas Landsat TM has four bands (TM 1-4) at VNIR, two bands (TM 5, TM 7) in the SWIRand one band (TM 6) in the TIR region. Depending up on the object (such as water, minerals/rocks, vegetation and soil) reflectanceand emittance of electromagnetic radiation will vary for different bands. The details of Landsat band and ASTER bands with spectral range and spatial resolution is given in the table 2.

Figure 3.Different channel of ASTER and Landsat with respect to the electromagnetic (EM) spectrum (Modified after Hajibapir et al. 2014). Wavelength of EM spectrum vs. reflectance of water, Al-OH mineral, vegetation and soil are also plotted for comparison.

 

4. Reflectance of different mineral in VNIR and SWIR

 

SWIR electromagnetic spectrum is very useful for identification of minerals and minerals group such as silica, clay, carbonates, iron oxides and other silicates. ASTER images have higher spectral resolution in the SWIR compare to Landsat; hence it is used for mineral identification. Complete reflectance and emittance spectra of different minerals / mineral groups for wide range of electromagnetic wavelength are experimentally generated and available at different library. In the figure 4 a) reflectance spectra (USGS library) of the studied minerals are superimposed on ASTER data band intervals in the VNIR-SWIR region(Elsaid, et al. 2014).In the figure 4.b) emittance spectra (Johns Hopkins University (JHU) library) of the studied minerals are superimposed on ASTER data band intervals in the TIR region (Elsaid, et al. 2014). This comparison helps us to use selected band for observing a mineral or group of minerals.

Figure 5. ASTER thermal band indicating how the emissivity of the rock changes with respect to silica content (Shift of minima to shorter wavelength as silica content increases) (Modified after Sabins, 1999).

 

Iron oxides, clay, carbonate and sulfide minerals show distinct absorption in the VNIR and SWIR regions. Iron oxide/hydroxide minerals (e.g. limonite, hematite, etc.) shows absorption from 0.4 to 1.1 μm of the electromagnetic spectrum due to the presence of transition elements such as Fe2+, Fe3+, Mn, Cr and Ni in the crystal structure (Table 3). Minerals bearing Al–O–H, Mg–O–H, Si–O–H, Al-Si-OH, Mg-Si-OH, Ca-Al-Si-(OH), other hydroxyl and CO3group (Clay mineral such as Kaolinite, montmorillonite, chlorite, illite; muscovite; talc; epidote group; carbonates such as calcite, dolomite; and sulphate minerals including alunite and gypsum) shows absorption (Table 3) in the SWIR region due to vibrational processes(Pour and Hashim, 2014; and reference therein). In the TIR region (between 8 – 14μm) quartz shows strong vibrational absorption due to asymmetric Si-O stretching (Pour and Hashim, 2014; and reference therein). Similarly carbonate minerals such as calcite and dolomite shows distinct emissivity absorption features are different wavelength of TIR region. These characteristics of minerals help to identify and map an area using remotely sensed images with different spectral range. In the figure 5 emissivity spectra of different igneous rocks are plotted for TIR region from 8 – 14 μm. Due to variation of the silica concentration in the rock center of the absorption band (broad emissivity minima in the spectra) change its position. In this figure center of the absorption band is indicated by an arrow and it shift to longer wavelength as the silica concentration in the rock decreases (Sabins, 1999).

 

 

Table 3: Absorption peaks of various elements/ions/ ionic group in different regions of EM (Gupta, 2003; Remote Sensing Applications, 2010)

 

4.1 Hyperspectral images

 

So far we have discussed about various multispectral sensors (e.g. ASTER, Landsat) having different bands in the VNIR, SWIR and TIR regions which can be used for mineral mapping, lithological discrimination. But using these data it is difficult to determine the detailed mineral composition and their relative abundance due to lack of finer band width and good spectral contiguity (Ramakrishnan and Bharti, 2015). However hyperspectral sensors acquire images from in contiguous (100-200) spectral bands with narrow bandwidth (10-15 nm spectral resolution). These images will provide a complete reflectance/ emittance spectrum. These spectrums are similar to the laboratory spectra and very useful for discrimination study to identify different mineral and rocks. The important airborne hyperspectral sensors used for mineral mapping and lithological discrimination are AVIRIS, HYDICE, DAIS, HyMAP, Hyperion and ALI etc. (Ramakrishnan and Bharti, 2015). In the figure 6, bands and images of hyperspectral and multispectral sensors are compared. The hyperspectral image acquisition is illustrated in the figure 7.

 

5. Characteristics of Images

 

All the remotely sensed images are processed before extracting any information about the object/material. Prior to image processing characteristics of image are needs to beknown(Whateley, 2006). The characteristics of mages are

 

  • No. of spectral band, swath, successive paths overlap, temporal resolution etc.
  • Pixel parameter: Digital images consist of discrete picture elements, or pixels. Each pixel associated with a number that signify the average radiance or brightness which represents reflected radiance from surface and radiation scattered by the atmosphere. Each is grouped as an 8 bit “byte” of information, with each bit used to indicate ascending powers of 2 from 20 (=1) to 27 (=128).
  • The Instantaneous Field of View (IFOV) is the distance between consecutive measurements of pixel radiance which is called as pixel size. For MSS: 79 m × 79 m, TM and ASTER has 30m x 30m pixel size.
  • Image parameter: Number of pixels per scene decides the resolution of images. In the multispectral images same scene is imaged simultaneously in several spectral bands. The image intensity level histogram is a useful indicator of image quality which describes the statistical distribution of intensity levels. (Whateley, 2006)

 

6.  Digital Image processing

 

For mineral resource study important function of digital image processing’s are

a) image restoration,

b) image enhancement, and

c) Information extraction.(Whateley, 2006)

 

a) Image Restoration: In this process images are corrected for inherent defects incorporated during data collection.

  1. Correction for lost data i.e. dropped lines or bad pixels
  2. Atmospheric noise correction
  3. Geometric correction

 

The earth rotates during the time it takes the satellite to scan its swath, resulting in the skewed image. Hence geometric correction is required as images are needs to be integrated with the geophysical,topographical or other map based data.

 

b)    Image enhancement:During this process original data of the images were transformed to the suitable form to improve the information content (Whateley, 2006). Important ones are:

  1. Contrast enhancement: Contrast stretch (a simple linear transformation) is routinely used to increase the contrast of a displayed image by expanding the original gray level range to fill the dynamic range of the display device.
  2. Spatial filtering: It is used to enhance linear surface features such as fractures, faults, joints, etc.
  3. Density slicing: It transformsthe continuous gray tone range into a series of density intervals (slices) with a specific digital range. Each slice may be given a separate colour or line printer symbol.
  4. False colour composite images (FCC): A multispectral false colour image is generated by combining different spectral band and assigning a different colour to each band. It increases the amount of information available for interpretation.

(c) Information extraction: After image restoration and enhancement, interpretation of images for extracting valuable information such as rock type, mineral deposits, alteration zone, structure etc. were carried out using spectral signature and other associated features. For image interpretation systematic observation of image elements and terrain elements is required (Remote SensingApplications, 2010) as each of these elements has their geological significance.

Image Elements: Tone/colour, Texture, Pattern, Shape, Size, Shadows, Site and Association

Terrain Elements: Drainage patterns (Table 4), drainage density, topography/land form and erosion status

 

Geological mapping and/or mineral exploration based on remotely sensed datarequired a) image interpretation of terrain element and image element with documentation of geological features (identification of features and judging their significance) based on the variation of spectral signature and b) verifying the result with the existing regional geological maps, reports, guides or by field visit. Generally broad geological information about the terrain is a prerequisite for preparing geological map or map for mineral exploration from air/space borne images. As discussed in the previous section for identification of different rock / mineral group/mineral their spectral signature is very helpful.

Other than the image element and terrain elements, geological information can be extracted from the image by various methods such as band ratio, spectral angel mapper, multispectral classification, principle component analysis etc.

 

 

  1. Band ratio:Gray level of pixel in one band divided a by another band. These ratio help to identify the ferruginous and limonitic capping which also called as gossan. Soils and exposed rocks rich in iron oxides and hydroxides absorb wavelength of <0.55 μm and cause red coloration (Whateley, 2006). And when these minerals mixed with other minerals it masks the coloration hence it is important to discriminate different minerals. Small contribution of iron minerals can be identified/ it can be enhanced by taking ratio of MSS band 4 over band 5. Similarly for discriminating areas of limonite alteration a ratio of MSS band 6 over band 7 will be helpful. Mineralogical spectral characteristics related to alteration are detected by taking ratio of TM bands (Whateley, 2006). ASTER provides better enhanced alteration discrimination due to finer spectral bands. Using this ratio we can also detect the alteration zone. For identifying hydrothermally altered rockLandsat TM bands 5 and 7 are very useful. A complete list of different band ratio (ASTER) used for identification of various mineral is given in the table 5.
  1. Spectral Angle Mapper (SAM): This method determines angle of similarity between the reference spectrum and an image spectrumby calculating spectral angle among them (Figure 8). Here image spectrum for unknown material is compared with reference spectrum of a known material; and spectra treated as multidimensional (n) vector while computing by the SAM. Where n is the number of bands. This method is widely applied for spectral matching and it is independent of illumination conditions (Lie et al., 2014) and albedo effects. Therefore it treats all possible illumination equally. The position of each spectrum under all possible illumination condition can be defined by a vector from the origin through each point. Poorly illuminated pixels are plot close to the origin. In the result if angle is small then it indicate the closer match with the reference spectrum. Regardless of the length the angle between the vectors is same. Under this method if the pixels are away from the specified angle (in radians) then they will not be classified. The direction of unit vector will define the color of a material (Exelis Visual Information Solutions, 2014).

Figure 8: Illustration of scatter plot of image spectrum and reference spectrum of two bands. Here t and r are two spectral vectors.

 

Table 5. Standard ASTER enhancements products. (Kaliknowski and Oliver, 2004, Source: ASL environment Sciences Inc. 2008)

 

3) Multispectral classification: In this process (digitally) a symbol or color is assign to a pixel or small group of pixels representing similar surface material. These selected pixels have high probability to represent same kind of material.To the large extent variation in the vegetation cover mimic the underlying geology, hence using multispectral classification this can be detected and the information about the geology can be extracted. For the large areamultispectral classification is also used to extract information. (Whateley, 2006).

 

4) Principal component analysis (PCA): Spectral variations among different rock type are more visible in the principal component image compare to single bands. Therefore principal componentanalysis helps to enhance or distinguish lithological differences. Very good correlation is observed in the reflectance of different bands of MSS, TM, or ASTER images. To improve the spread and to exaggerate differences in the data principal componentanalysis is applied by redistributing data on another set of axes (Whateley, 2006).

 

Flow charts describing the steps generally followed for preparing/updating lithological map (Figure 9) and integrating remote sensing data with various geoscientific data for mineral exploration purpose (Figure 10) are provided below.

 

Figure 9. Flow Chart describing steps for preparation / updation of lithological map from satellite image.(Source: Remote Sensing Applications, 2010,NRSC / ISRO, Hyderabad, India)

Figure 10.Steps used in integration of remote sensing data with other geoscientific data for mineral exploration. (Source: Remote Sensing Applications, 2010, NRSC / ISRO, Hyderabad, India)

 

7. Case study:

 

To learn how this technique can be used for mineral exploration, an example is given here. This case study is directly quoted from “ Elsaid, et al. (2014): Processing of Multispectral ASTER Data for Mapping Alteration Minerals Zones: As an Aid for Uranium Exploration in Elmissikat – Eleridiya Granites, Central Eastern Desert, Egypt, The Open Geology Journal, 8, (Suppl. 1: M5) 69-83”. Here author tried to map the alteration mineralization and/or zones as pathfinders for uranium mineralization within Elmissikat – Eleridiya younger granite (Central Eastern Desert, Egypt), (Figure A 1.a) which may assist in developing a uranium exploration program using ASTER imagery. With above objective author used ASTER image followed by ortho-rectification using SRTM topographic data. Author also converted the pixel radiance to the reflectance at surface data using FLAASH (fast line of sight atmospheric analysis of spectral hypercubes) module of ENVI software and thermal atmospheric correction of ASTER-TIR emittance bands were also applied.False color composite of ASTER bands (9:R, 8:G, 1:B) was prepared to discriminates the different rock units of the study area (Figure A 1.b). The feature oriented principal component selection (FPCS) and spectral angle mapper (SAM) is the main image processing techniques was used by the author. The input bands for FPCS analysis of selected minerals (type of minerals will vary as per the aim of the study) are listed in the table A 1 below. Format for result of principal component analysis for selected bands (VNIR-SWIR and TIR separately) and for a mineral is given in the table A 2. For the selected minerals (as per the table A1) separate analysis is required. Based on the result anomaly maps will be prepared for each selected minerals. Few maps are given in the figure A 2 a, b,c,d). For SAM classification computed maximum angle in radians are listed in the table A 3 for selected endmember minerals. Prepared map based on the SAM analysis is given in the figure A 3 a) and b).

 

 

Table A 1: Input bands for FPCS (Feature Oriented Principal Component Selection) analysis of the selected minerals (Elsaid, et al. 2014).

Table A 2: Principal component analysis and Eigenvector values for Selected VNIR-SWIR ASTER Bands for Kaolinite (upward arrow means reflectance and downward ones mean absorbance) for the study area is listed down(Elsaid, et al. 2014). Results of principal component analysis are not shown in the table here, as it only for the demonstration purpose. Band DN (Digital Number) Threshold = Mean + 2*(Standard Dev.)

Table A 3: Matching Maximum Angles (Radians) used in SAM classification for ASTER Bands

used for the study area (Elsaid, et al. 2014).

FigureA 1. a) Geologic map of the study area, Elmissikat – Eleridiya district, Central Eastern Desert, Egypt(Elsaid, et al. 2014; and reference therein),Blue circle indicate the uranium exploratory mine. b) False color composite of ASTER bands (9:R, 8:G, 1:B) discriminates the different rock units in the study area. (Elsaid, et al. 2014)

Figure A 2. a)Image for Kaolinite based on the PC3 of FPCS technique for the input VNIR-SWIR bands applied on the study area. Magenta pixels represent kaolinite anomalies. (Elsaid, et al. 2014)

 

b) Image for Kaolinite based on the PC3 of FPCS technique for the input thermal emissivity bands applied on on the study area. Magenta pixels represent thresholdedkaolinite

 

c) Image forIllite based on the PC4 of FPCS technique for the selected input VNIR-SWIR bands applied on the study area. Blue pixels represent thresholdedillite(Elsaid, et al. 2014)

 

d) Image forHematite based on the PC1 of FPCS technique for the input VNIR-SWIR bands applied onthe study area. Red pixels represent Hematite (Elsaid, et al. 2014)

 

e) Image forQuartz based on the PC5 of FPCS technique for the input thermal emissivity bands applied on the study area. Cyan pixels represent threshold quartz (Elsaid, et al. 2014)

 

Figure A3.a) separated SAM classification for end-member (Quartz) using ASTER VNIR-SWIR stack and the USGS ASTER resampled spectral library applied on the study area, b) separated SAM classification end-member (Quartz) using ASTER TIRstack and the USGS ASTER resampled spectral library applied on the study area. (Elsaid, et al. 2014)

 

Result from the above study reported by the author is: From FPCS and SAM image the resulted alteration minerals maps (especially silica) nearly coincide with radioactive anomalies and could be used as pathfinders for new uranium exploration sites within the study area. For more information and interpretation please refer Elsaid, et al. 2014.

 

8 Conclusion:

 

This module introduces the importance of mineral resource study and an example was given that for a person how much mineral, metal or stone required in his or her lifetime. Here we also learned how remote sensing and GIS techniques can be used for mineral exploration, preparing/ updating lithological map of an area. Using reflectance and/or absorbance of different wavelength of electromagnetic radiation, minerals/rocks can be identified and this property of mineral will also help us to use remote sensing technique more efficiently in the various fields of mineral exploration and mapping. Usefulness of remotely sensed images from different sensor is also discussed here to guide or select suitable image for mineral resource study. In this module only selected digital image processing and image interpretation techniques are discussed as detailed discussion is beyond the scope of this module. The case study presented here is only to demonstrate that how one can use this technique and what are the steps involved for mineral exploration study.

 

9.  References:

  • ASTER Mineral Exploration, ASL Environment Sciences Inc., 2008.
  • E-Book on Mineral Sector, Ministry of Mines, Government of India, February 08, 2016
  • Elsaid M., Hatem Aboelkhair, Ahmed Dardier, Elsayed Hermas and Urai Minoru, (2014): Processing of Multispectral ASTER Data for Mapping Alteration Minerals Zones: As an Aid for Uranium Exploration in Elmissikat – Eleridiya Granites, Central Eastern Desert, Egypt, The Open Geology Journal, 8, (Suppl 1: M5) 69-83.
  •  Exelis Visual Information Solutions, 2014
  •  Gupta RP, 2003, Remote Sensing Geology, Springer-Verlag,2nd Ed., p 498-524.
  • Hajibapir G., Mohammad Lotfi, Afshar Zia Zarifi, Nima Nezafati (2014): Application of Different Image Processing Techniques on Aster and ETM+ Images for Exploration of Hydrothermal Alteration Associated with Copper Mineralizations Mapping Kehdolan Area (Eastern Azarbaijan Province-Iran), Open Journal of Geology, 2014, 4, 582-597
  • Harris J. R., J. Peter, L. Wickert, P.H. White, M. Maloley, R. Gauthier and P. Budkewitsch, 2014: Hyperspectral Remote Sensing for Geological Mapping and Mineral Exploration- A Review of Activities at NRCAN, Natural Resource Canada.
  • Kanlinowski, A. and Oliver, S., 2004. ASTER Mineral Index Processing. Remote Sensing Application Geoscience Australia.
  • Liu L., Jun Zhou, Fang Yin, Min Feng, Bing Zhang , (2014): The Reconnaissance of Mineral Resources through ASTER Data-Based Image Processing, Interpreting and Ground Inspection in the Jiafushaersu Area, West Junggar, China. Journal of Earth Science, Vol. 25, No. 2, p.397–406.
  • Palomera R. P. A., 2004, ‘Application of remote sensing and geographic information systems for mineral predictive mapping, Deseado Massif, Southern Argentina’ M.Sc. thesis, submitted to International Institute for geo-information science and Earth observation, Enschede, The Netherlands.
  • Pour A.B. and Hashim M., (2014): ASTER, ALI and Hyperion sensors data forlithological mapping and ore minerals exploration, , Springer Plus, 3:130
  • Ramakrishnan D. and Bharti R., (2015): Hyperspectral remote sensing and geological applications, Current Science, vol. 108 (5).
  • Remote Sensing Applications, 2010, NRSC / ISRO, Hyderabad, India
  • Sabins F. F. (1999), Remote sensing for mineral exploration. Ore Geology Reviews 14. 157– 183
  • Whateley M. K.G. (2006). “Remote Sensing”. Introduction to mineral exploration. 2nd ed. / edited by Charles J. Moon, Michael K.G. Whateley& Anthony M. Evans, Blackwell Publishing, USA.

Web resources: