•  
  •  
 

Abstract

Image segmentation is one of main and necessary goals in image processing for digital image, it seeks to segment the studied images into multiple useful regions that summarizing interest regions for satellite images, which are multispectral images equipped by satellites using the concept of remote sensing, which has become an important concepts that it's applications used in most of the necessities of daily life, especially after the rapid developments in various fields of life that many of these fields have used algorithmsand software techniques, these images are very necessary to enable us to study a wide range of goals in many scientific fields, in this research, the nonhierarchical cluster analysis algorithm was used as a method of image segmentation (splitting and merging regions) in order to demonstrate the importance of using statistical methods in image processing tasks, such as image segmentation, where (K-Means) technique was used to implement this task, this algorithm was applied on multispectral satellite image of a scene from western Iraq, where the results showed the flexibility of this algorithm in dealing with the disparity in the lighting of color image pixels and it's efficiency of formation clustering region that composed from groups of homogeneous pixels in there degree of illumination intensity, finally, the ability of this algorithm to give a good quality images which are measured using peak signal to noise ratio (PSNR) scale to measure image quality.

DOI

10.33095/jeas.v25i111.1638

Subject Area

Statistical

First Page

466

Last Page

484

Download

Share

COinS