Research On Coal Quality Index Testing Method Based On Data Analyses Of Coal Images

At present, coal quality index testing tools commonly used thermal analysistechniques to analyze and test the coal, although high accuracy but time-poor, usuallyderived from coal into the plant to test results takes2-3days, which is not beneficial tobusiness blending coal, coal into the furnace of coal price calculation, accounting costsand so on. Rapid detection of coal quality index has become urgent problems to thermalpower enterprise.On the basis of a detailed analysis of coal images and its corresponding coal qualityindex, propose the method of coal quality index detection based on big data analyses ofcoal images, which uses image recognition technology and includes coal imagepreprocessing, image feature extraction and selection and classification decision. Coalimage preprocessing module uses digital image processing techniques to achievesegmentation of the whole coal samples from coal images. Image feature extraction andselection module extracts the image color and texture features, in which color featuresinclude color moments and color histogram, and texture features include GLCM andTamura, then chooses image feature which best reflect coal quality situation of coalimages. Classification decision module bases on minimum distance classification, whichmeasures the distance between the image features by Euclidean distance.Experimental results show that coal quality index testing method based on dataanalyses of coal images is feasible; Selecting color histogram and Tamura texture featureto test coal quality index has higher accuracy rate. Coal quality index testing has a certaindeviation because coal images are gray and have small amount of color information.