| Summary: | The Malmberget iron ore deposits is located in the Norrbotten province in Sweden. The iron ores of Norbotten are characterized by intensive recrystallization during different metamorphic events and by the presence of both magnetite and hematite as the main ore minerals. Currently geological and mineralogical information is obtained mainly through visual observation of drill cores and interpreting chemical analyses done on selected intervals. Proper mineralogical techniques like optical microscopy, X-ray diffraction and scanning electron microscopy are used only in special mineralogical studies mainly in the development stage but not any more when mining operation has been started.The aim of this study is to evaluate the state-of-the-art drill core scanning technique: hyperspectral imaging (SisuROCK and SisuCHEMA) that has been developed and is already used in some mines for ore characterization. The technique is not capable for identifying iron oxides, but also many of minerals which were expected to be recognized with hyperspectral imaging could not be positively identified. In addition study aimed to develop a novel method to collect a mineral map of hematite-magnetite ore directly from drill cores for textural analysis.As a result two potential methods were found, further developed and tested. Fingerprint technique uses iron powder which sticks on magnetite and a fingerprint on the distribution of magnetite can be imaged. Combining this information with optical image can give an image where magnetite, hematite and gangue minerals can be separated and information can be used for textural analysis. IR Thermography techniques utilise thermal camera and difference in the thermal properties of minerals present in iron ores, potentially also magnetite and hematite can be positively identified.The fingerprint technique and IR Thermography are promising due to the ability in distinguishing magnetite and hematite. Further improvement is needed in order to make fingerprint and IR Thermography technique become automated ...
|
|---|