Study Of Coal Spontaneous Combustion Of The Coal Mine Monitoring Method

The coal is one of the important energy in our country, and it has a vital role in the sustainable development of national economy. But coal is also one of the most dangerous industries in the industrial production. Mine fire is a major highlight hazards, as its uncertainty of harm, the fire, caused by spontaneous combustion of coal, is much greater.For the early characteristics of coal spontaneous combustion, there are a variety of detection methods, such as temperature measurement method, the method of index gases, smell method, resistivity exploration method, isotope method for measuring radon concentration and so on. Each method has its advantages and disadvantages, but the single detection method is difficult to achieve for a more accurate monitoring of coal spontaneous combustion condition, and detecting informations are not intuitive, this topic proposed will be applied to the image recognition technology, infrared image was recorded using infrared thermal imager, the image can vivid reflect the temperature change of the spontaneous combustion. The image characteristics, as a kind of detection, by understanding the change of the high temperature area of infrared thermal image, using multi-sensor information fusion technology of information fusion. By infrared thermal image to identify the spontaneous combustion of coal.First of all, this paper introduces the principle of infrared thermal imaging, to get rid of the noise from the image, which collected by the infrared thermal imager. This subject adopts threshold denoising based on Contourlet transform, improved threshold function, has obtained the good effect in getting rid of the noise. Then, to extract the edge of infrared image, using the improved Sobel edge detection algorithm, can effectively detect the edge of infrared image information, edges are clear and continuous.Secondly, as single detection method can’t reflect the degree of coal spontaneous combustion in time, put forward a comprehensive evaluation model of multi-sensor information fusion, to judge the degree of coal spontaneous combustion.Finally, aiming at the uncertainty of multi-sensor information, puts forward the fuzzy C-means clustering method based on genetic algorithm, to classify the sensor information collected. For coal spontaneous combustion monitoring in a large number of complex fuzzy qualitative events, put forward the combined neural network and fuzzy reasoning algorithm, the fuzzy inference system based on neural network is established. Through the establishment of the inference rules, toidentify the collected sensor information of the nature. Through the training simulation, it has very good effect in identifying the degree of coal spontaneous combustion.