A synergistic PCA and C-A fractal approach for high-precision extraction of mineral alteration anomalies from ASTER remote sensing data

Abstract

Listen

Translate article

Traditional remote sensing algorithms struggle to effectively identify altered minerals. This study proposes a method that integrates ASTER image eigenvalue principal component analysis (PCA) with the concentration-area (C-A) fractal model to extract multiple types of alteration anomalies in mining areas. PCA is first applied to extract key feature components, minimizing background noise from surrounding land cover. Subsequently, the C-A fractal method quantifies alteration intensity, utilizing a grid-based segmentation and change-point model to classify alteration levels and distinguish them from other geological features. This approach refines anomaly detection, overcoming the limitations of traditional techniques and improving classification accuracy. A case study in the Shizhai mining area of Guangdong indicates that the extracted iron-staining, hydroxyl and skarn alteration distributions align closely with known favorable mineralization zones. Additionally, this method enables the precise extraction of chlorite and epidote, providing critical geological insights for mineral exploration.