Automatic image recognition system and key technologies of maceral group

Coalification degree and macerals determine the physical and chemical properties as well as coking proper- ties of coal. The automatic determination of vitrinite reflectance and macerals in coal can reduce the deviation from tra- ditional manual determination. Also,it is quick and efficient,which makes it possible for the determination results of coal petrology to be used in coal classification and the processing and utilization of coal. Numerous research has been done worldwide in the automatic determination of coal petrology. The automatic determination of vitrinite reflectance has been achieved. However,due to the complexity of maceral composition and the influence of coalification degree, there are still many difficulties in the automatic recognition of macerals and image analysis. In this paper,the research focuses on these problems. The following results have been obtained:① the microscope-based hardware platform for micro-image automatic collection is developed,which includes three functional modules:autofocus,automatic scanning and automatic collection of micro-image. The workflow of automatic image recognition of maceral group is established. ② Micro-image de-noising pretreatment techniques are developed for segmenting resin from liptinite group and the brightened resin affected by reflection of coal particles below. Based on Prewitt operator,a removal technique of false boundary of maceral is also developed. ③ A K-means clustering-based image automatic segmentation and recognition technology of maceral groups is developed. ④ Automatic image recognition system of maceral group of bituminous coal is developed. Using the image automatic recognition system developed in this research,the representative samples of bituminous coal from different ranks in China are automatically collected and their macerals are automatically recog- nized. The results are compared with the standard results of manual identification done by domestic authoritative ex- perts in coal petrology field. It is shown that the average deviation of the 45 samples in vitrinite,inertinite and liptinite group determined by the two methods are only 2. 3% ,2. 3% and 1. 5% respectively. According to the accuracy analy- sis method in national standard GB / T 18510—2001,the statistic tc is less than the tt of three maceral groups.