Astronomical Image Clustering Using Spectral, K-Means and Mean Shift

By analyzing vast amounts of observational data through astronomical image analysis, you can gain valuable information about the structure of the universe. This paper presents a new method for clustering astronomical images that uses both spectral and K-means methods and also incorporates linear algebra and optimization principles. We propose a method that combines feature extraction techniques such as the Fourier transform with clustering algorithms to classify celestial objects based on their characteristics. The research examines the effectiveness of spectral clustering and K-means clustering in organizing astronomical images into meaningful groups, taking into account factors such as computer efficiency, clustering accuracy and interpretability. Experimenting with different datasets, including telescope images, we demonstrate the effectiveness of our approach in automating the classification of astronomical objects. Our results reveal insights into the underlying structure of astronomical knowledge and offer promising avenues for future research in computational astronomy. In general, this research advances both the field of astronomy and the application of mathematical methods to image analysis and clustering.