| Coal is an organic rock. Coal petrology is a subject which uses coal-rock approach to research coal. With the development of coal chemical technology, the application of coal petrology theory and practice increased breadth and depth in the field of coke. Coal maceral indexes mainly include the quantitative statistics of coal macerals and the measurement of vitrinite reflectance.Combining these two indicators, it can control the coal quality, evaluate the coal coking properties exactly and predict the coke strength, optimize the blending options to improve the coke quality and use coal economically and reasonably. Currently, the major equipment of coal petrographic analysis is microscope photometer which is composed by the microscope and spectrophotometer microscope photometer. It can meet the test requirements of the indicators, but there are also problems. On the one hand, maceral identification is also totally dependent on manual, so the entire testing process is time-consuming.On the other hand, it required a series of hardware devices, such as a spectrophotometer, high voltage power supply, so the system cost is high. Therefore, only a very small number of units have the equipment, not widely used.Based on the gray value determination results of organic macerals obtained: first, on the same coal, the inertinite, vitrinite, exinite gray value distribution components are in descending order, while on the different coals, three gray level is the same with the increase in metamorphic grade. Second, vitrinite reflectance is in fact image brightness under certain light. Based on this conclusion, the paper proposed to use image processing methods to achieve the quantitative statistics of coal macerals and the measurement of vitrinite reflectance. Compared with traditional system, the system based on digital image analysis will accelerate the pace of coal analysis, reduce human error, reduce product cost, lay the foundation for the promotion of the system and the rational use of coke resources.The quantitative statistics of coal macerals based on digital image processing are mainly through the count of each component to achieve. To make the count accurate, the initial images must be pretreatment. Image denoising experiments show that wavelet transform for image denoising is better than mean filtering, no significant difference compared with median filter. Taking into account of future embedded applications, this paper uses median filter for image denoising. Binarization processing adopts the method of automatic search for the best threshold value. This method automatically analyzes the image histogram, according to the three main components of coal to determine the optimal threshold value, and then uses the optimal threshold for binarization processing. After comparing several color image edge sharpening algorithms, including Robert operator, Prewitt operator, Sobel operator, Canny operator, LOG operator, finally adopts the LOG operator, and has achieved a good edge enhancement. After this series of treatment macerals binary image can be gotten.Maceral statistics can be given by counting the processing image.The measurement of vitrinite reflectance based on digital image processing are mainly through histogram analysis of I value to achieve. HSI model has the characteristics which can separate brightness information I from the color of the image. After analyzing the gray value of specific maceral each point, formates the histogram. The histogram shows more intuitive coal vitrinite reflectance characteristics.Based on the above analysis, the article gives coal analysis system's hardware structure and software algorithm, and this system is based on spatial digital image processing. Compared with the the traditional analysis system, this new system reduced the processing time from a few hours to a few minutes, reduced the costs from nearly a million to less than one hundred thousand. |
PDF Full Text Request