Morphometric classification of lunar domes using unsupervised clustering and orthogonal descriptors

Abstract

Lunar domes record signals of both surface and subsurface volcanic and thermal processes; therefore, a robust morphometric classification facilitates the interpretation of their genesis and provides insights into the thermal evolution of the lunar crust. In this study, we propose a new classification based on quantifiable morphological descriptors derived from digital elevation models (DEMs), morphometric parameters, and image analysis. Unlike previous approaches, this method does not rely on theoretical estimates that may introduce uncertainty into classifications. We analyzed 26 domes, including the 16 intrusive domes compiled by Lena et al. (2013) and a representative selection of 10 effusive domes. The classification was based on geometric properties (slope, height, diameter, volume, and area) and Jacobi–Fourier Descriptors (JFDs) extracted from grayscale images. Unsupervised clustering methods (K-means and Agglomerative Clustering) were applied to evaluate 26 combinations with varying JFD complexities and numbers of clusters. Cluster quality was assessed using internal validation indices (Silhouette, Calinski–Harabasz, and Davies–Bouldin). The best performance was achieved with Agglomerative Clustering using 2×2 JFD descriptors and k = 3, yielding consistent internal metrics (Silhouette = 0.4280; CH = 17.06; DB = 0.8581). Reduced pairs of JFD moments (e.g., |J_(2,2)1,1| and |J_(2,2)1,2|) demonstrated strong discriminative power among volcanic styles. The resulting clusters partially reproduced the classification proposed by Lena et al. (2013) but revealed intrusive morphological subtypes not previously recognized, suggesting that surface morphometry contains relevant signals regarding emplacement mechanisms. This methodological approach provides an objective and replicable alternative for classifying landforms without relying on theoretical models or spectral/compositional information, and it can be applied to other planetary bodies or volcanic landforms to support volcanic and thermal evolution comparative studies.