Comparative Assessment of Image Fusion Methods for Land Cover/ Land Use Monitoring
R. Prema1, M. G Sumithra2
1R.Prema, Assistant Professor, Department of ECE, Bannari Amman Institute of Technology, Sathyamangalam (Tamil Nadu), India.
2M.G Sumithra, Professor, Department of ECE, Bannari Amman Institute of Technology, Sathyamangalam (Tamil Nadu), India.
Manuscript received on 07 December 2018 | Revised Manuscript received on 18 December 2018 | Manuscript published on 30 December 2018 | PP: 105-110 | Volume-8 Issue-2C2, December 2018 | Retrieval Number: 100.1/ijeat.B10231282C218/18©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Land Cover/Land Use (LCLU) applications includes rural/urban change detection, biomass mapping and natural resource management. The high spatial and high spectral resolution are necessary for efficient class discrimination of LCLU monitoring and mapping. The multispectral (MS) images acquired from land observation satellites like Landsat, MODIS and IRS etc. are able to provide only coarse spatial resolution. The Panchromatic (PAN) band in satellite images have high spatial but coarse spectral resolution. The process of combining PAN and MS band images to produce high spatial and spectral resolution is called Image Fusion. Classical image fusion algorithms are Brovey, Intensity Hue and Saturation (IHS), Principal Component Analysis (PCA), High Pass Filtering (HPF), Atrous Wavelet Transform (ATWT) and Generalized Laplacian Pyramid (GLP). These benchmarking methods are coming under pixel level fusion. In this paper, we are going to analyze the performance quantitatively using evaluation parameters Spectral Angle Mapper (SAM), Universal Image Quality Index (UIQI) and Relative Global-dimensional Synthesis Error (ERGAS). The experiment is performed using datasets Landsat and QuikBird. All the simulations were carried out in MATLAB R2014a. Comparison of all the methods concludes the better approach for future research.
Keywords: Image Fusion, Land Cover/Land Use, Pixel Level Fusion, Spatial Resolution and Spectral Resolution.
Scope of the Article: Image analysis and Processing