Miner fuzzy detection based on mixture Gaussian model in underground coal mine videos

In order to monitor the dangerous areas in the underground coal mine automatically, a miner fuzzy detection method based on mixture Gaussian model was proposed in this paper. Recent history of each pixel was modeled by a mixture of K Gaussian distributions; the mean value of every Gaussian distribution constructed the background images; the difference images and the frame image were transformed into the fuzzy field. Some fuzzy rulers were defined according to the characters of coal mine videos, then the fuzzy system got every pixel´s membership value in the object to classify it as object or not. The proposed method was compared with segmenting difference image based on 2-D fuzzy entropy and background subtraction based on mixture Gaussian model. The experimental results show that this method can remove the interference of miner´s lamp and detect the miner effectively even they are similar to the background. This method has the characteristics of less calculation; it is suited for practical use in underground coal mines.