Computer Science · Artificial Intelligence
Advanced mineralogical classification and concentration estimation in mining with MATLAB-powered hyper-spectral imaging and machine learning
Natsuo Okada, Brian Bino Sinaice, Jae-Won Kim, Hiromasa Nozaki, Kaito Takizawa, Narihiro Owada, Yoko Ohtomo, Youhei Kawamura
See on the mapAbstract
The study presents a new technique that combines hyperspectral imaging and machine learning to identify minerals. Overcoming the challenges of processing hyperspectral data, the solution offers a user-friendly hyperspectral analysis tool tailored for constructing datasets and enhancing mineral identification and concentration estimation. The tool integrates hyperspectral data processing with segmentation, simplifying complex operations and making mineral identification accessible to non-experts. The tool’s capabilities extend to handling multi-spectral data efficiently, potentially leading to energy-efficient analysis when combined with dimensionality reduction techniques. Presented as a novel approach, it improves geological surveys in mining areas, enabling industrial applications and mineral research.