Characterization of Combustion Chemiluminescence: An Image Processing Approach

The applicability of image data processing to estimate the spatial distribution of Air/Fuel ratio at the flame front on gas combustion systems are presented and discussed. The objective is to devise a procedure capable of converting a single RGB image data, obtained by conventional CCD cameras, into a reliable sensing on local combustion state for practical application on atmospheric flames of premixed methane and propane gases. In this context, this paper proposes a fast and innovative image processing methodology. An image processing algorithm based on the camera sensor transfer function, a lookup calibration table, and sparse matrices operations is proposed. The optimized version of our image processing algorithm is able to process high resolution images at rates as high as 0.6 Mpixel/s on a standard personal computer (3.5 GHz clock frequency). As a result, color RGB images are converted into quasi real-time detectors of local combustion state for practical flames of CH4 and C3H8 gases. It has been experimentally observed that camera blue and green channels mimic the flame CH* and C* chemiluminescence emissions, revealing a wide applicability on φ detection (Air/Fuel ratio), roughly ranging between 0.8 and 1.4.