Automated vein detection for drill core analysis by fusion of hyperspectral and visible image data

Abstract

The analysis of veins is crucial for understanding the genesis of many mineral deposit styles, providing important information for exploration and mining companies to vector towards economic base or precious metal deposits. However, even though automated vein detection is fundamental for the collection of objective quantitative results that can be implemented in resource models, related published literature is limited. While image analysis can be potentially used to segment veins from a drill core image, it may not always work due to the complexity of the drill core image. When the color and texture of a vein is not obviously contrasted to its background in a drill core image, the vein may not be detected at all. This paper presents a data fusion based approach for mapping quartz and carbonate veins in drill cores. The proposed approach uses different image analysis techniques together with the abundance data of minerals produced from the unmixing of the thermal infrared (TIR) and short-wave infrared (SWIR) spectra. The experimental results show that the proposed data fusion method successfully logged the quartz and carbonate veins and has laid the groundwork for further development to ‘paint’ the results onto drill core imagery.