Automated mineralogical anomaly detection using a categorization of optical maturity trend at lunar surface.

The mineralogical anomaly is the mineralogically diagnostic character that differs from its surrounding spectra in terms of absorption features of a spectrum. On the airless planetary body like our Moon, the characteristics of mineral exposures are mostly affected by the physical process over time because of space weathering. This inevitable fact produces mature exposures that act as a background for mineralogical anomaly detection. In this regard, the present study categorizes the optical maturation trend of a site, and the background characterization is integrated with anomaly detection, which makes it more practical, robust, and efficient. These fundamental ideas have been improvised into our automated mineralogical anomaly detection method that uses a modified density-based spatial clustering of applications with noise (Modified DBSCAN) algorithm to categorize the maturity trend from various scatter plots of spectral property. Subsequently, the Reed-Xiaoli detector analyzes the data points belonging to below 33% of maximum point density on the corresponding scatter plot. Experimentally, the proposed method has been validated with previously reported results of literatures over the varietal lunar sites (crater Theophilus, crater Humboldt, Hansteen Alpha silicic dome, western inner ring of Mare Moscoviense, and crater Von Kármán), which have diverse mineralogy in the region. The reflectance spectra of anomalous points explore the mineralogical diversity of the studied region using Chandrayaaan-1 Moon Mineralogy Mapper ( M 3 ) data. However, it can also operate with any hyperspectral datasets within the spectral range of 400–3000 nm. This flexibility makes our proposition more attractive for an end-user in the field of planetary science. [ABSTRACT FROM AUTHOR]