Scholarlink

This study presents a prediction of titanium dioxide (TiO \(_2\) ) concentration utilizing advanced machine learning techniques and spectral analysis. To train a model based on the Multivariate Adaptive Regression Splines (MARS) technique, we integrated remote sensing data from the Clementine mission, compositional analyses of lunar samples from the Apollo missions, and laboratory reflectance spectra from the RELAB database. The spectral consistency between Clementine and RELAB datasets enabled a robust correlation for TiO \(_2\) estimation. To provide an independent and external validation, a separate dataset of 34 lunar samples analyzed with data from the Moon Mineralogy Mapper (M3) instrument aboard the Chandrayaan-1 mission was employed. This integrated approach enhances the accuracy of TiO \(_2\) distribution mapping by combining spectral algorithms validated with in-situ sample analysis, laboratory spectroscopy, and independent M3 data. Spatial analysis using geospatial tools enabled the visualization of TiO \(_2\) distribution across diverse geological contexts, including lunar maria and impact craters. This study highlights the importance of integrating spectrophotometric techniques with machine learning models for lunar mineralogical mapping, offering new insights into lunar regolith evolution and enhancing future exploration and resource utilization efforts.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Prediction of Titanium Dioxide (TiO \(_2\) ) Concentration Using Machine Learning and Spectral Analysis

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Julia Fernández Díaz,
  • Francisco Javier de Cos Juez,
  • Fernando Sánchez Lasheras,
  • Javier Gracia Rodriguez,
  • Santiago Iglesias,
  • Sara Rodríguez

摘要

This study presents a prediction of titanium dioxide (TiO \(_2\) ) concentration utilizing advanced machine learning techniques and spectral analysis. To train a model based on the Multivariate Adaptive Regression Splines (MARS) technique, we integrated remote sensing data from the Clementine mission, compositional analyses of lunar samples from the Apollo missions, and laboratory reflectance spectra from the RELAB database. The spectral consistency between Clementine and RELAB datasets enabled a robust correlation for TiO \(_2\) estimation. To provide an independent and external validation, a separate dataset of 34 lunar samples analyzed with data from the Moon Mineralogy Mapper (M3) instrument aboard the Chandrayaan-1 mission was employed. This integrated approach enhances the accuracy of TiO \(_2\) distribution mapping by combining spectral algorithms validated with in-situ sample analysis, laboratory spectroscopy, and independent M3 data. Spatial analysis using geospatial tools enabled the visualization of TiO \(_2\) distribution across diverse geological contexts, including lunar maria and impact craters. This study highlights the importance of integrating spectrophotometric techniques with machine learning models for lunar mineralogical mapping, offering new insights into lunar regolith evolution and enhancing future exploration and resource utilization efforts.