A significant amount of metallic iron in the world is produced in blast furnaces from iron ore pellets. To control the quality of pellets and to optimise pellet production, characterisation of both green and fired pellets is required. Optical image analysis (OIA) allows identification of different minerals, binder, remaining fluxes and porosity within pellets.
For proper identification of the mineralogy, porosity and texture in pellets a high magnification is required, therefore each individual image covers only a very small portion of the pellet. At the same time the characterisation of pellets requires that the whole pellet is represented as a single object, so the size of the resulting overall image at high magnification can make it difficult to handle. A solution to this problem has been implemented within the CSIRO OIA package Mineral4 by combining images at different levels. Namely, to average out local irregularities, individual frames can be stitched during imaging into larger MosaiX images. Mineral4 further combines these MosaiX images into a single panorama image of the whole pellet.
The proposed procedures allow the calculation of mineral and porosity spatial distributions, local gradients and abundance graphs based on the distance from the centre of the pellet. The ‘Pellets’ module of CSIRO OIA software also allows for characterisation of large irregularly shaped particles of lump ore and sinter.
Poliakov, A, Donskoi, E, Hapugoda, S and Lu, L, 2017. Optical image analysis of iron ore pellets and lumps using CSIRO software Mineral/Recognition, in Proceedings Iron Ore 2017, pp 583–592 (The Australasian Institute of Mining and Metallurgy: Melbourne).