الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 128 |  |  |
| | | from | 128 | | |
Abstract
Fundamentally, remote sensing is the process of getting information about the earth without making a physical contact with the earth surface. Currently, remote sensing has become extremely important for many applications; namely, hydrology, ecology, oceanography, glaciology, geology. Furthermore, it is also used in military, commercial, intelligence, humanity applications. Basically, there are two types of images which can be obtained from remote sensing. The first type is the panchromatic image with high spatial resolution and low spectral resolution; on the other hand, the second one is the multispectral image which has low spatial resolution and high spectral resolution. Due to the importance of remote sensing, several applications require images with high spatial resolution and high spectral resolution. The next paragraph illustrates the definition of image fusion and its techniques. As a general rule, image fusion is the process of merging the two previous types of images the panchromatic image and the multispectral image to get in the final one image with high spatial resolution and high spectral resolution. Actually, there are different methods of image fusion. Basically, the methods applied in this research are Intensity Hue Saturation (IHS), Fast IHS with spectral adjustment, fast IHS with spectral adjustment area coefficient, high pass filter (HPF), hyperspectral color, principal component analysis (PCA), Gram Schmidt, Brovey, Wavelet + IHS, and Wavelet+PCA. The first step in image fusion processing is to resize the pixel size of the multispectral image to be equal to the pixel size of the panchromatic image and this process is called image resampling. There are three techniques of image resampling. These techniques are, namely, the nearest neighbor method (nn), bilinear interpolation method (bi), and cubic convolution (cc). Therefore, this research aims to study the effect of the three image resampling techniques on the different image fusion methods to see whether image resampling affects image fusion or not. One of the most important applications in remote sensing is image classification. Image classification is the process of classifying and identifying the features in the image. Because of image fusion, there would be a change in the digital number of the pixel and consequently there would be a change in some spatial or spectral information. Thus, this research aims to study also the effect of the applied image fusion techniques in this research on image classification to determine which method of the previous image fusion methods renders the best, the highest, and the acceptable overall accuracy. There are different techniques in image classification Maximum Likelihood, neural network, and minimum distance. The method employed in that point is the maximum likelihood classification method.