Research On Color Image Salt And Pepper Noise Filtering Method Based On Uniform Spatial Color Difference Metric

In the industrial converter steelmaking, steelmaking predictive analysis of data is very important, a direct impact on steel production efficiency and product quality. An important part of the image information as a data converter steelmaking predictive analysis, providing an important basis for the data to predict due to aging industrial field presence of soot and electronic devices and other reasons easy to form salt and pepper noise, damage to the image information is larger, It will therefore be carried out late segmentation, image feature extraction, image processing chain cause greater interference, thereby affecting the validity and accuracy of the data converter steelmaking prediction, based on the above background, this paper, salt and pepper noise color image filtering approaches were studied.Conventional color image filtering algorithms are mainly concentrated in the RGB color space, RGB space and non-uniform space, and keep the texture edge color and other details of the deficiencies, industrial flame image for more attention to the actual situation of the texture color images and other information, this paper presents a color image based on impulse noise filtering algorithm uniform color space measured around the algorithm completed the following work:Based on the type of noise, color space and image quality evaluation criteria analysis, by color space conversion image RGB space under the transition to the next CIEL* a* b* space, based on the study of color uniform space metrics, calculated according to the color difference formula proposed by the international Commission on illumination and CIEDE2000 were CIE1976 algorithm analysis and implementation. And algorithms for CIEDE2000 noise density at high-performing issue, the optimization algorithm based on adaptive window, and analyzed comparative experiments.To further improve the color of impulse noise filtering effect of high noise density, uniform color space based on the analysis of the distribution of the proposed color image filtering algorithm based on uniform color space multi-directional measurement, the results show that, in maintaining good image textures and colors at the same time, the algorithm can more effectively remove the noise, the image used in subsequent research work to provide a quality assurance.