A New Method of Image Denoising for Furnace Flame

摘要

Furnace flame image mainly contain impulse noise and Gaussian noise. For this feature, the denoise method of median filter and wavelet transform combined are proposed. The specific method is: Firstly, remove the impulsive noise for furnace flame with adaptive weight median filter, and then decompose the image with Daubechies. At last more effective filtering are used in different sub-bands, it adopt different approaches against the characteristics of the different sub-bands. Experimental results show that the proposed method can effectively remove the impulse noise and Gaussian noise and improve the quality of the flame image, and it is better than the traditional denoising method.