Coal Images Database and Its Application for Selected Coal Grains Properties Analysis

The raw materials processing methods play important role in excavation industry. The area is nowadays a target for novel, innovative technology application. One of them is an image processing as an instrument for determination of different material properties. The paper proposes a tool – database system – dedicated for supporting images processing and computer vision methods application in the mining industry. The idea and implementation of the system is presented. The database was supplied with a set of coal images. The images were taken at a special stand, with the calibrated camera. The usefulness of system was tested by application of a pipeline of image processing algorithms. The goal was to estimate the grain size composition of coal and rock matter presented on images. Two algorithms for edges detection were tested: Sobel filter and Statistical Dominance Algorithm. After edges detection, a set of morphological operations were applied of determination of good starting markers for the watershed algorithm. The result showed that both algorithms tend to produce over segmented images. However, SDA performed noticeably better. The experiments proved the usefulness of developed database system dedicated to image processing.