Iron Ore Tailing Composition Estimation Using Fused Visible–Near Infrared and Thermal Infrared Spectra by Outer Product Analysis

Iron ore tailings are mainly composed of SiO2 and iron, whose content determines the potential reuse strategy tailings. Compared with traditional wet chemistry approach, spectroscopy has proven its superior effectiveness in characterizing predicting minerals, such as iron oxides, clay, SiO2. This study aims to estimate TFe based on visible–near infrared (VIS–NIR, 350–2500 nm) thermal (TIR, 8–14 μm) spectroscopy. The outer product analysis (OPA) method is used combine VIS–NIR TIR spectral domains, from which an matrix fusion data can be generated. area tailing dam Waitoushan, one super-large deposits Anshan–Benxi cluster northeastern China. results demonstrated following: (1) reflectance feature at 1163–2499 nm range correlates emissivity 8–9.4 10.7–12 μm (2) original absorbance spectra, correlation coefficients spectra improve 0.66 0.87 for 0.64 0.84 (3) partial least squares regression, random forest (RF), extreme learning machine exploiting particle swarm optimization modeling methods established estimation. prediction accuracy indicate that model OPA-fused performs significantly better than individual spectra. RF input-fused provides highest determination 0.95 0.91, root mean square errors 0.97% 0.96%, ratios performance interquartile distance 6.49 2.31 estimation, respectively. These outcomes provide a theoretical basis technical support composition estimation using