Clearing Research On Fog And Dust Images In Coalmine Video Surveillance System

Coal is one of the most important resources in our country, but it has some difficulty in mining coal,and some accidents occur easily in the process. The cause of accidents has two parts: on the one hand is lacking of awareness, the other hand is backward technology and machine. Safety is the hottest topic in the society, so in the process of coal exploitation,we should pay more attention to personal safety.This paper studies clearing research on fog and dust images in coalmine video surveillance. It can reduce the coal accidents by raising the quality of the image, prevent the accidents,and is helpful to rescue victims.Clearing research on fog and dust images in coalmine video surveillance mainly has two factors(fog dust and noise), in order to get clear images, it must process two factors.In denoising process, firstly,it described two common noises in detail: salt and pepper noise and gaussian noise; secondly, it introduced two kinds of noise filter methods: median filter, bilateral filter and their related algorithm; Finally, expounded the advantages and disadvantages of two methods.In fog and dust, according to the dark-channel method proposed by Dr He Kaiming, we compared the improved algorithm with the old one, and summarized its advantages and disadvantages.In order to clear coal fog dust images with random noise, first of all,the paper introduced an algorithm of fog dust removal and simultaneously denoising based on DCPBF, the principle: it establishes a coalmine fog dust image degradation model based on fog model according to the model,the dark-channel method is suitable for the coalmine fog dust image degradation model, the fine transmittance diagram was obtained by bilateral filtering, on the basis of building the coalmine fog dust image degradation model, we calculate the transformation image and process fog dust and noise image by gaussian bilateral filtering.An adaptive image algorithm was putted forward for poor quality image in the video monitoring system. Self-adaption mainly reflected in the estimated value of A light atmosphere. The principle : firstly, it is different from common methods that getting the A_max of the dark channel image; secondly, according to the different average gradient value of the images,concluded its suitable estimates of the atmospheric light(the proportion of A_max) by statistical methods; finally, recovered image by A and the algorithm of fog dust removal and simultaneously denoising based on DCPBF(the old algorithm) and got the new algorithm, the new algorithm was better than the old algorithm,and improved the reliability of the old algorithm.