1. Introduction
Nowadays, remote sensing is considered as an important subject in many field studies, like agronomy, hydrogeology, mineral exploration, geography and geology. This knowledge has been applied in most of the geological fields. The significant application of remote sensing is preparation of lineament maps that is an important issue in the geological studies. In the remote sensing study, geological structures (e.g. cracks, fractures, faults, shear zones and foliations) are illustrated as a lineaments (Kamali et al.; 1390,2). Therefore, tectonic and structures of different areas can be investigated with use of remote sensing techniques and providing these maps.
Lack of access to some areas as well as necessity to spend a lot of money and times for studying large areas have been forced researchers to find a solution for these problems. The first remote sensing studies related to geological applications dates back to first aerial photo studies in the early 20thcentury (Koç; 2005, 6). The lithology, structures and geology units had been identified with using aerial photographs, and finally geological maps had been drawn. After the Landsat images were published in 1972, the use of remote sensing was increased in the earth studying. Recently these images are largely applied in the geological researches. Since satellite images are obtained from varying wavelength intervals of the electromagnetic spectrum, they are considered to be a better tool to discriminate the lineaments and to produce better information than conventional aerial photographs (Casas et al.; 2000, 2011). The purpose of this research is to apply the remote sensing techniques and ASTER images for lineament extraction (faults) using both visual (manual) and automated methods in the central Alborz, map faults preparation, and then comparison of these maps with geology maps and with together.
2. Study Area
Iran is located within the Alpine–Himalayan folded belt in a zone of continental convergence between Eurasia and Arabia platforms which is mostly accommodated by shortening and strike-slip faulting in the mountain belts such as the Great Caucasus, Zagros, Alborz, KopehDagh and also the active Makransubduction zone (Javidfakhr et al.; 2011, 290). The Alborz range is an active mountain belt that surrounds the South Caspian Basin. The deformation of the Alborz is due to the northward shortening between the central Iranian block and the Eurasian plate (Vernant et al.; 2004, 179).
The researches show that the present-day deformation in Alborz is characterized by range-parallel left-lateral strike-slip and thrust faults (Djamour et al.; 2010, 1). The study area is located in the northern part of the Central Alborz Mountain range between longitude of 51.19˚-52.13˚E and latitude of 36.01˚-36.68˚N.The main structures of the study area are the North Alborz and Khazar faults, which are only separate structures east of 51°30'E. The North Alborz Fault is mapped as a south-dipping thrust (Geological Survey of Iran 1991a, b, Allen et al.; 2003, 662). It changes strike at 53° E from ENE to WNW. The Khazar Faultwith a southward dip direction is the longest fault in the northern edge of the Alborz Mountains. This structure separates the Caspian Sea to the north from the Alborz Mountains in the south (Motaghi et al.; 2010, 790).
3. Material and Methods
In this study, three data sets have been used:
1) ASTER Digital Elevation Models(DEMs) of central Alborz (between longitudes of 51.19˚-52.13˚E and latitudes of 36.01˚-36.68˚N).
2) 1:25000 contour line maps of central Alborz.
3) 1:100000 geological maps of Baladeh and Marzan Abad
Before the lineament extraction, a set of image processing techniques should be done that are including radiometric and geometric images correction and image enhancement techniques. Since the radiometric correction had been performed in the ground stations on the satellite images, this correction wasn’t necessary for our ASTER images. The geometry correction and geo-referencing have been done only on VNIR images due to be high spatial resolution. The image enhancement techniques that applied on VNIR band are including filtering (directional), PCA (principle component analysis) and band rationing (OSAVI (Rondeaux et al.; 1996, 102)). Then, lineaments have been extracted using both visual (manual) and automated digital interpretation method.
4. Results and Discussion
After lineament extraction the fault maps, rose diagram, density and length-density maps have been drawn. The comparison analysis of obtained faults maps by both methods (visual and automated)with faults maps that obtained from geology maps was done in points of estimating frequency, length, direction and density of line aments. The frequency of automatically extracted lineaments (3886) are greater than the manually ones (1912). The first reason for this results is that short lineaments have been identified by the program (LINE module of PCI Geomatica), and the second one is that this program can’t distinguish faults from other lineaments (roads, rivers, etc.). The amount of identified lineaments with filtering, PCA and band rationing are respectively 1440,411 and 57. As you see, the frequency of identified lineaments with filtering and PCA are greater than the band rationing ones. The reason is due to further lineaments enhancement using directional filters as well as 80% information in the first component of the PCA.With using band rationing some of the lineaments in the southern part of the study area which could not be recognized by directional filters and PCA have been identified. The average length of the manually extracted lineaments (2.19 km) that are close to the geology map lineaments (1.93) are far from the average length in the automatically extracted ones (0.18). The maximum length of lineaments for the manually extracted (25 km) which is as long as fault lengths in the central Alborz (e.g. 150 km for Mosha fault) is 0.36 km in the automatically ones. Orientations of lineaments for both maps are illustrated with rose diagrams. The dominant orientations for the manually extracted lineaments are in the east-west and northwest-southeast directions. These orientations are as a result of deformation in the region. Dominant deformation in Alborz is characterized by range-parallel left-lateral strike-slip and thrust faults. Deformation is due to the north–south Arabia–Eurasia convergence, and westward motion of the adjacent South Caspian relative to Iran. Roughly north–south shortening occurs on thrusts that dip inwards from the range margins (Allen et al.; 2003, 659). The density (Zakir et. al.; 1999, 1073) and length-density maps indicate increasing of lineament density in the northern parts and in east-west direction (in the direction of Khazar and North Alborz faults). This means that there is a fracture zone in the area (Sarp; 2005, 58).
5. Conclusion
In this research, lineaments (faults) of central Alborz are extracted from the ASTER images by using remote sensing techniques and both visual (manual) and automated methods. The results show that directional filters and PCA had a main role in faults identification and extraction.The reason is due to further lineaments enhancement using directional filters as well as 80% information in the first component of the PCA. Some of the lineaments in the southern part have been identified by using band rationing. The extracted lineaments are consistent with the geology maps’ faults (Saidi; andGhassemi; 1381; VahdatiDaneshmand, 1379). These images processing techniques and lineaments identification indicate that directional filters (Sarp; 2005, 43. Koçal; 2004, 15) are the most important factors for lineaments recognition in the vegetated areas.Frequency, length and orientation of the automatically extracted lineaments have no consistency with the geology maps’ faults. Since the program (PCI geomatica) can’t distinguish fault lineaments from the others, the Khazar and North Alborz faults which are the most important structure in the Alborz region have not been recognized. So, for geological researches that always have been based on field studies and require high precision, the automated method because of the low accuracy isn’t recommended. In the visual (manually) method, length, orientation, density and length density are consistent with fault’s status in central Alborz. Consequently, the best way for geological and structural studies on large areas or areas that are difficult to access, is the visual method for lineament extraction. |
