The power generation industry relies heavily on coal despite the availability of other energy sources. The use of low quality coals, and coal blends from a variety of sources is becoming widespread in power plant for economic and availability reasons. Co-firing coal with biomass on existing coal fired furnaces is recognised as one of the new technologies for reducing CO2 emissions in the UK and the rest of the world. The changes in these fuel supplies have posed significant technical challenges for combustion plant operators and engineers to maintain high combustion efficiency and low atmospheric emissions including CO2, NOx, SOx and particulates. Despite various advances in developing the coal combustion and co-firing technologies, a range of technological issues remain to be resolved due to the inherent differences in the physical and combustion properties between coal and biomass. A typical problem associated with the use of low quality coal and co-firing of coal and biomass is the uncertainty in the combustion characteristics of the fuels, often resulting in poor flame stability, low thermal efficiency, high pollutant emissions, and other operational problems. To meet the stringent standards on energy saving and pollutant emissions, advanced technology for improved understanding of energy conversion, pollutant formation processes and consequent combustion optimisation in coal-biomass fired furnaces have therefore become indispensable.A flame, as the primary zone of the highly exothermic reactions of burning fuels, contains important information relating closely to the quality of the combustion process. Recent study has shown that the combustion process, particularly the pollutant emission formation processes, can be better understood and consequently optimised by monitoring and quantifying radical emissions within the flame zone through spectroscopic imaging and image processing techniques. It is proposed to develop a methodolgy for the monitoring and quantification of the radiative characteristics of free radicals (e.g. OH*, CH*, CN* and C2) within a coal-biomass flame and consquently the estimation of the emission levels in flue gas (e.g. NOx, CO2 and unburnt carbon). A vision-based instrumentation system, capable of detecting the radiative characteristics of the multiple radicals simultaneously and two-dimensionally, will be constructed. Computing algorithms will be developed to analyse the images and quantify the radiative characteristics of the radicals based on advanced signal processing techniques including wavelet analysis. The relationships between the characteristics of the radicals and fuel type and air supplies will be established. The emission levels in flue gas will be estimated based on characteristic features of the flame radicals obtained by the system. All data processing will be performed in an industrial computer system associating with integrated system software including a graphic user-interface. The system developed will be initially tested on a gas-fired combustion rig in University of Kent and then an industrial-scale coal combustion test facility run by RWE npower. A range of combustion conditions will be created during the industrial tests, including different coal-biomass blends and different fuel/air flowrates. The relationships between the emission characteristics of radicals and the chemical/physical properties of the fuels and the pollutant emissions will then examined under realistic industrial conditions.The outcome of this research will provide a foundation for a new area within coal-biomass combustion optimisation in which advanced flame monitoring techniques could help to predict emissions directly from the flame information instead of the flue gas measurement, shortening the control loop for emissions reduction. Such techniques would greatly benefit the power industry by allowing them burning fuels more efficiently and meanwhile reducing harmful emissions to the environment.