Fractional-Order Differentiate Adaptive Algorithm for Identifying Coal Dust Image Denoising

Due to the complex underground mine environment, it is very difficult to process the images with large amounts of coal dust noise. So fractional-order differential modeling with adaptive algorithm is introduced to eliminate image noise. This paper applies firstly the theory of fractional calculus based on the AA model (Aubert-Aujol model)to model for image processing, and then according to the regional characteristics, selects model parameters -the fractional order u and regularized parameter λ of each image point by means of the adaptive algorithm. Numerical experiments show that the quantitative indicators to measure noise effect-the peak signal-to-noise ratio (PSNR) and edge-preserving index (EPI) are better based on the improved algorithm than the traditional. Because of the good denoising effect in the "non-texture region" and a good texture retention capacity in the "texture region", Fractional-order differential adaptive algorithm is a fast and efficient image denoising method. In conclusion, this new algorithm is applied to detect coal dust in underground and achieves satisfactory results.