Novel developments in optical image analysis for iron ore,sinter and coke characterisation

Abstract

A textural approach to the geometallurgical characterisation of iron ores helps better predict ore behaviour during downstream processing. Therefore, a robust, automated, objective method for the textural characterisation of iron ores is relevant to both research and industry needs. Utilisation of an optical image analysis (OIA) technique allows reliable and consistent identification of different iron oxide and oxy-hydroxide minerals, e.g. haematite, kenomagnetite, hydrohematite, both vitreous and ochreous goethite. CSIRO Mineral4/Recognition4 OIA system automatically identifies particle sections with different textures and assigns these sections to defined textural groups. Furthermore, novel developments in the system have enabled the automatic identification of different textural forms and morphologies of the same mineral, e.g. martite and microplaty haematite in iron ore; primary and secondary haematite or different types of Silico-Ferrite of Calcium and Aluminium (SFCA) in sinter as well as segmentation of different phases with the same reflectivity like Inert Maceral Derived Components (IMDC) from Reactive Maceral Derived Components (RMDC) in coke.

The high resolution and imaging speed of the OIA system makes it possible for users to significantly reduce the cost and subjectivity of iron ore characterisation with a simultaneous increase in the accuracy of mineral identification. Extra software modules have been developed to meet research and industry demands for enhanced productivity. The addition of a ‘Multiple Block Imaging’ module enables image acquisition for sets of polished blocks at a time, rather than separate imaging of individual blocks. The ‘Multiple Set Processing’ module allows the processing up to 20 groups of sets simultaneously, where every group can contain up to 20 different sets of images that share the same analysis profile. The added modules enable analyses to be performed over many hours without the need for operator intervention, thus increasing equipment utilisation and reducing operator time.

These new developments, together with the improvement of previously available features, e.g. identification of non-opaque minerals, automated textural classification, automated particle separation, automated correction of mineral maps and on-line measurement, means that OIA can provide a unique, reliable, industry and research focused tool for iron ore, sinter and coke characterisation.