Integrating Texture Features with a Weighted Differential Evolution Algorithm for Satellite Image Clustering

Satellite image data classification is crucial for categorizing pixels into meaningful classes. While traditional reliance on spectral information has limitations, a shift to textural features from the co-occurrence matrix proves essential. This paper introduces the Weighted Differential Evolution Algorithm (WDEA) for numerical optimization, focusing on CHI, XB, and DBI indices. The evaluation includes Partitioning Around Medoids (PAM) and Hierarchical Centroid (HC) methods. Comparisons with the K-means (KM) algorithm confirm the effectiveness of WDEA in satellite image clustering.