Image fusion remote sensing for the extraction of urban features / Ahmad Nadzari Yahaya

• Main Author: Yahaya, Ahmad Nadzari
• Format: Article
• Language: English
• Published: Unit of Research, Development and Commercialization (URDC),UiTM Perlis 2004
• Subjects: Applied optics. Photonics; Photoelectronic devices (General)
• Online Access: http://ir.uitm.edu.my/id/eprint/11547/1/AJ_AHMAD%20NADZARI%20SABRI%20YAHAYA%20JI%2004.pdf; http://ir.uitm.edu.my/id/eprint/11547/

Rapid advances in computer image analysis have allowed for greater flexibility and the use of techniques for combining and integrating multiresolution and multispectral data. With the multiresolution and multispectral satellites the fusion of image data has become a valuable tool in remote sensing image evaluation. This study evaluated several methods to enhance the spatial resolution of multispectral images using a higher resolution panchromatic image. This was performed using Landsat7 ETM+Multispectral bands and Pancromatic band imagery satellite. The Landsat7 ETM+ panchromatic band is taken simultaneously with multispectral bands using the same system, therefore, are co-registered accurately. The study involved image and theme enhancement by applying Optimum Index Factor(OIF) using three techniques of image fusion : Intensity-Hue-Saturation(IHS), Brovey and Principal Component Analysis(PCA). Brovey was observed to be the best fusion technique for extraction urban features with respect to less spatial and spectral distortion as well as visualization. OIF is a powerful technique for the analysis and visualization of urban features. Based on training samples, ETM+ multispectral and fused images were subjected to the process of Iterative Self-Organizing Data Analysis (ISODATA) techniques using maximum likelihood decision rule to derive the classification of urban features classes. Signature separability using transformed divergence was used to evaluate between signature classes for classification. Ground truth for classification accuracy assessment were evaluated using error matrix and overall classification techniques. This would display the advantages and results of fused images classification.