Research On Fracture Detection Algorithm Of Coal Rock Image

In recent years, energy shortage has become the focus of the world attention. Thedevelopment of high-quality energy has drawn attentions from each country. Our country isrich in coalbed methane resources, ranking third in the world. When we are mining, it willdischarge more than13billion cubic meters of coalbed methane every year. Coalbed methanehas the same calorific value with natural gas, and they can be mixed to transport and use. Thecombustion of coalbed methane is pure, virtually no emissions. There are a wide range ofusage of coalbed methane, such as domestic fuel, industrial fuel, fuel for power generation,automotive fuel and an important chemical raw material. Coalbed methane commonly knownas “gas” and if we do not timely exploit coalbed methane when we are mining, it may lead togas explosion. When the density of coalbed methane in the air up to5%-16%, in case of firewill explode, at the same time, if the coalbed methane emitted directly into the atmosphere, itsgreenhouse effect of carbon dioxide is about21times, devastating the ecological environment.So the development of coalbed methane has great significance: firstly, it can improve theenergy structure, increase clean gas energy; secondly, it can improve coal mine safetyproduction environment, improve economic efficiency; thirdly, it can effectively reducegreenhouse gas emissions, improve the atmospheric environment. Coal rock fracture detectionis the key way to develop coalbed methane. Accurately detect the fracture of the coal rockimage plays an important role in the exploitation of coalbed methane. The main innovations inthe dissertation are outlined as following:1) An algorithm of combining fuzzy enhancement with fuzzy mathematical morphologyfor coal rock image fracture detection is presented. Due to the coal rock image has too manynoise and the complexity of coal rock image background, we must enhance the image beforewe extract the fracture edge. Two kinds of fuzzy enhancement--adaptive fuzzy enhancementbased on slide window and multi-level fuzzy enhancement are used to widen the gray value offracture edges and background, and make it convenient for the next fracture edge detection.Fuzzy morphology can effectively suppress noise, and accurately detect the coal rock fracture.Experimental results show this algorithm can effectively detect and extract the fractureinformation of coal rock image, and suppress the noise.2) A method of combining multi-structure elements morphology with self-organizingmap neural network for coal rock image fracture detection is presented. Multi-structureelements morphology edge detection algorithm are noise immunity, at the same time, it candetect all the edges of coal rock image. Because of the complexity of coal rock imagebackground, we not only get the fracture edges but also we get non-fracture edges.Self-organizing map neural network is self-organized and unsupervised. Unlike the BP neuralnetwork need to specify the output results. Firstly, the detected binary image are regionlabeled, and then the feature parameters of coal rock edges are calculated as the inputs ofself-organizing map neural network. Finally, the fracture edges and the non-fracture edges areclassified by the self-organizing map neural network clustering algorithm, and the fractureedges of image are gained. Experimental results show this new method can effectively detect and extract the fracture information of coal rock image.