Experimental verification of three-dimensional temperature field reconstruction method based on Lucy-Richardson and nearest neighbor filtering joint deconvolution algorithm for flame light field imaging

Abstract

Non-intrusive flame detection technique, such as radiation imaging, can measure large flame in boilers or engines. Flame light field imaging is a new type flame detection method based on radiation imaging. Using this method, a multi-view flame light field image can be obtained in a single capture, and the three-dimensional temperature field of the flame can be calculated using a reconstruction algorithm. In this study, experiments are conducted to verify the temperature reconstruction method using the Lucy–Richardson and nearest neighbor filtering joint deconvolution algorithm based on the refocused flame light field images. It was found that, in terms of the temperature trend and range, the results for different flame sections can be reconstructed to be comparable to the results using the non-negative least squares method. Although the flame shapes are non-axisymmetric corresponding to working conditions, the temperature reconstruction method using refocused light field images can effectively reflect these changes.