Flame temperature estimation from light field image processing.

ABSTRACT

We propose a novel flame temperature estimation method based on a flame light field sectioned imaging model of complex temperature distribution in different media. The proposed method relies on multi-pixel reconstruction to improve the resolution of sub-aperture images. In addition, the wavelet transform denoises the flame refocused image, and then the Lucy-Richardson algorithm deconvolves the image. The temperature estimation accuracy using the proposed method is higher than that reported in a previous work, with a larger temperature estimation range from 1250 K to 1800 K. Moreover, we found that deconvolution plays an important role in determining the temperature estimation accuracy.