The segmentation of aerial images using different clustering techniques offers valuable insights into interpreting and analyzing such images. By partitioning the images into meaningful regions, clustering techniques help identify and differentiate various objects and areas of interest, facilitating various applications, including urban planning, environmental monitoring, and disaster management. This paper aims to segment color aerial images to provide a means of organizing and understanding the visual information contained within the image for various applications and research purposes. It is also important to look into and compare the basic workings of three popular clustering algorithms: K-Medoids, Fuzzy C-Mean (FCM), and Gaussian Mixture Model (GMM). This will help find the best way to separate colors in aerial images. According to a thorough comparative study, PSNR and correlation metrics show that K-Medoids outperform other clustering techniques in terms of segmentation quality. Also, the effect of changing the number of clusters on the image quality was studied; when the number of clusters increases, the image quality increases. It was found that when K-Medoids were used, the PSNR and correlation were 35.57 and 0.99, respectively. When FCM and GMM were used, they were 35.54, 0.99, 31.67, and 0.97, respectively, when the number of clusters was 12.
