Automated Image Analysis System for Coal Petrographic Characterisation

A low-cost imaging system has been developed to characterise coal by accurately determining the reflectance distribution of its constituents. It is a tool that can be used by coal producers, coal users, and commercial and research laboratories. It can be fitted to existing petrographic microscopes that are used for manual coal analyses. It can provide a characteristic reflectance 'fingerprint' of any coal and a commercially valid rank measure for a wide range of single seam coals. The output is more detailed and more consistent than that obtained from manual methods alone. One of the authors, who is not a trained petrographer, has gained an international accreditation for coal petrography using the system. This is the first time a computerised system has been used successfully to meet an international coal petrography standard.

This work was performed under a continuation of ACARP project C6055. A prototype system was developed in the first phase of the project that was reported in 1998 (O'Brien et al, 1998). Recent advances within the system include:

• Topography correction software to compensate for inconsistencies in specimen preparation.
• The ability to process grains individually.
• Ability to analyse a wider range of coals in terms of rank has been significantly increased.
• The system is able to quantify microbreciation.
• The system is able to analyse raw coals and can separately quantify the abundances of bright (pyrite) and dark (clay) minerals.
• The system is able to discriminate between dark minerals and liptinite. Further work is required to refine this capability.
• Significant component cost reductions.

A feature of the system design is that it produces images of coal that are calibrated for reflectance at the individual picture element level. These images are suitable for further processing, for example, to provide analyses of blended coals or to provide measures that can be used in performance models such as coal grain combustion. Thus, the system affords an opportunity to develop further new tools to assist coal research and enhance the value of light microscopy techniques.

The current phase of development is complete. Further significant improvements can only be realised by tailoring the system to perform specific industrial applications and testing its performance under operating conditions. To assist in this next stage of development and commercialisation the software has been licensed by CSIRO to Jenkins-Kwan Technology Pty Ltd.