Conference Proceedings
ICAM 2008 - Ninth International Congress for Applied Mineralogy
Conference Proceedings
ICAM 2008 - Ninth International Congress for Applied Mineralogy
Qualitative and Quantitative Evaluation of Iron Ore Sinters Through Digital Microscopy
Vale (previously CVRD - Companhia Vale do Rio Doce) evaluates the microstructure of sinters in order to understand their characteristics and behaviour in the agglomeration process, and the correlation with conventional quality parameters. Efforts on image analysis development are in progress between Vale and PUC-Rio in order to improve both the qualitative and quantitative evaluation of a representative portion of material. In the present work, digital microscopy techniques were developed allowing this kind of assessment._x000D_
Employing a motorised and computer controlled optical microscope, low magnification mosaic images covering complete sinter samples were automatically obtained. The mosaics provide a powerful qualitative overview of the whole sample while allowing low resolution quantitative evaluation of the major phases. In addition, quantitative results for phase fractions were compared for the same sample regions at three magnifications (5x, 10x and 20x objective lenses). A single field (1x 1) at the lowest magnification corresponds to mosaics of 2x 2 and 4x 4 images at the higher magnifications. The automated system allows perfect registration between the mosaics and the single image. Changes in phase discrimination were evaluated as a function of optical resolution, and phase fractions obtained by interactive segmentation of the image histograms were compared. While finer dark phases, eg silicates, were only visible with the best optical resolution, coarser phases such as haematite and pores were measured at all resolutions. For haematite, magnetite, ferrite and pores, the largest relative error in area fraction, between the images obtained with the 5x and 20x lenses, was 12 per cent. The results for silicate were less accurate with a relative error of 44 per cent. The results indicate that digital microscopy provides a flexible method for qualitative and quantitative evaluation of these materials.
Employing a motorised and computer controlled optical microscope, low magnification mosaic images covering complete sinter samples were automatically obtained. The mosaics provide a powerful qualitative overview of the whole sample while allowing low resolution quantitative evaluation of the major phases. In addition, quantitative results for phase fractions were compared for the same sample regions at three magnifications (5x, 10x and 20x objective lenses). A single field (1x 1) at the lowest magnification corresponds to mosaics of 2x 2 and 4x 4 images at the higher magnifications. The automated system allows perfect registration between the mosaics and the single image. Changes in phase discrimination were evaluated as a function of optical resolution, and phase fractions obtained by interactive segmentation of the image histograms were compared. While finer dark phases, eg silicates, were only visible with the best optical resolution, coarser phases such as haematite and pores were measured at all resolutions. For haematite, magnetite, ferrite and pores, the largest relative error in area fraction, between the images obtained with the 5x and 20x lenses, was 12 per cent. The results for silicate were less accurate with a relative error of 44 per cent. The results indicate that digital microscopy provides a flexible method for qualitative and quantitative evaluation of these materials.
Contributor(s):
J C Alvarez, D T Wagner, G Schinazi, O da Fonseca Martins Gomes, M H de Pinho Mauricio, S Paciornik, M B Vieira
-
Qualitative and Quantitative Evaluation of Iron Ore Sinters Through Digital MicroscopyPDFThis product is exclusive to Digital library subscription
-
Qualitative and Quantitative Evaluation of Iron Ore Sinters Through Digital MicroscopyPDFNormal price $22.00Member price from $0.00
Fees above are GST inclusive
PD Hours
Approved activity
- Published: 2008
- PDF Size: 0.719 Mb.
- Unique ID: P200808094