ORIGINAL PAPER Application of multifractal modeling for the identification of alteration zones and major faults based on ETM + multispectral data Ramin Aramesh Asl & Peyman Afzal & Ahmad Adib & Amir Bijan Yasrebi Received: 20 December 2013 /Accepted: 5 March 2014 # Saudi Society for Geosciences 2014 Abstract The aim of this study is to investigate the recon- naissance of alteration zones and faults in Hashtjin 1:100,000 sheet (NW Iran) using concentration-area (C-A) fractal model based on remote sensing data, which has been extracted from enhanced thematic mapper (ETM)+ multispectral images. There are Oligocene volcano-plutonic rocks and Tarom- Hashtjin metallogenic zone with Cu, Au, and Pb-Zn occur- rences in the studied area. The concentration-area (C-A) frac- tal model proposed in this paper for the interpretation of pixel value distribution spatial patterns based on the extracted data from ETM+ multispectral images. The pixel values were calculated by the PCA (principal component analysis) method for iron oxides and argillic alteration. Furthermore, the sharpen-filtering has been applied to calculate the value pixels for the main fault zone in the Hashtjin area. The C-A model can be used to establish power-law relationships between the area and the pixel value. The log-log C-A plots show multifractal nature for iron oxides, argillic alteration zones, and faults. Results obtained by the fractal model reveal that alteration zones and major faults have a NNW-SSE trend. The alteration zones and major faults have a strong correlation with the geological map of the area. Keywords Concentration-area (C-A) fractal model . PCA . Multispectral . Hashtjin . Iran Introduction Multispectral images have been increasingly utilized to char- acterize features on the earth’ s surface for various purposes especially geosciences and mineral exploration. Remote sens- ing techniques are used for mineral exploration and geosciences in two applications: (1) geological mapping of faults, fractures, and lineaments; (2) delineation of hydrother- mal alteration zones (Beiranvandpour and Hashim 2012; Khan et al. 2007; Sabins 1999). The purpose of displaying the multispectral image should not only be to provide a visual representation of the variance of images, although this has been the primary objective of most conventional methods. The color palette should reflect real-world features on the ground which must be the primary objective of employing remote sensing data. One of the main tasks involved in image processing is to classify image values into components and to establish the relationships between these components and features on the surface. These relationships can be visualized with a proper visual illustration or analyzed by means of various quantitative methods. Nevertheless, the first view of the image with an appropriate color palette can be important since it gives users the first perspective of the spatial distribu- tion of multispectral image pixel values (Cheng and Li 2002). Principal component analysis (PCA) is a common method of analysis for correlated multivariable datasets, and the tech- nique is widely used for multispectral image interpretation based on linear algebraic matrix operations. PCA can effec- tively concentrate the maximum information of many corre- lated image spectral bands into a few uncorrelated principal components and therefore can reduce the size of a dataset and enable effective image RGB display of its information. This links to the statistical methods for band selection that aims to select optimum band triplets with minimal inter-band R. A. Asl : P. Afzal : A. Adib Department of Mining Engineering, Faculty of Engineering, South Tehran Branch, Islamic Azad University, Tehran, Iran P. Afzal (*) : A. B. Yasrebi Camborne School of Mines, University of Exeter, Penryn, UK e-mail: P_Afzal@azad.ac.ir Arab J Geosci DOI 10.1007/s12517-014-1366-2