Research On Key Technology Of Coal And Gangue Intelligent Recognition In Fully Mechanized Caving Face

Fully mechanized top coal mining has become one of the main methods of thick coal mining in china due to its high-yield and high-efficiency characteristics.The "unmanned" or "less humanized" of fully mechanized caving face is of great significance for achieving safe and efficient production in coal mines.And the intelligent recognition of coal gangue is a necessary condition to realize the intelligentization of fully mechanized caving mining.At present,in mechanized top coal caving mining,the operator judges the coal caving situation through hearing and visual inspection.When it is judged that the coal caving stream contains more than a certain proportion of gangue,the tail beam and the sliding plate are manually controlled to stop the coal caving.The above-mentioned coal caving process completely relies on the experience of the workers,which often results in that the coal on the top of the support is not discharged or the coal caving contains a large amount of gangue,resulting in low coal recovery rate or many impurities.Manual control of coal caving has high labor intensity,low work efficiency,high randomness,and low intelligence,which cannot meet the current construction needs of intelligent fully mechanized caving face.Therefore,it is necessary to study the key technology of intelligent recognition of coal gangue in fully mechanized caving face,and then improve the intelligent level of fully mechanized caving mining.In this paper,the coal gangue image is used as the recognition basis,and the accurate calculation of the gangue visual ratio is the goal.The gangue image is enhanced and segmented,and the key feature space that can characterize the surface characteristics of coal and gangue is extracted.Through the improved extreme learning machine classifier,coal gangue classification is optimized,and the calculation of coal gangue visual ratio is further realized.The main work and research results of the paper are as follows:(1)Combined with the main equipment and functions of the fully mechanized caving face,the comprehensive mechanized top coal mining technology is analyzed.The coal gangue image taken by the industrial camera is selected as the basis,and the functional requirements of the gangue intelligent recognition system of the fully mechanized caving face are analyzed,and a fully mechanized caving face is built.The overall architecture of the coal gangue intelligent identification system at the working face is built,and the coal gangue intelligent identification process is designed.(2)The main characteristics of coal gangue images in the transportation system of fully mechanized caving mining face are studied.Aiming at the problems of halo artifacts,over-enhancement and noise amplification in dark areas in the traditional Retinex algorithm in processing coal gangue images,an improved multi-scale Retinex image enhancement algorithm based on HSV color space and two-dimensional empirical mode decomposition is proposed,and using test images and field images,a simulation comparison analysis was carried out to verify the effectiveness and applicability of the improved algorithm.(3)Combining the improved quantum genetic algorithm and the optimal entropy threshold segmentation algorithm,the optimal entropy threshold coal gangue image segmentation method based on the improved quantum genetic algorithm is designed,the preliminary segmentation of the coal gangue image is completed,and the further segmentation of bonded ore is realized based on the watershed algorithm.Using the chain code and line segment table to extract the feature space of the ore monomer,through the improved extreme learning machine to complete the classification of the coal and gangue monomer,and further proposed the calculation method of the gangue visual ratio.(4)A coal gangue intelligent identification experimental platform is designed and built,and conducted simulation experiments and industrial experiments in the collaborative innovation center of the province and the ministry of mining intelligent mining equipment and the Gengcun Coal Mine of Henan Dayou Energy Co.,Ltd.The experimental results show that the method proposed in this paper can effectively improve the quality of coal gangue images and realize the intelligent recognition of coal gangue in fully mechanized caving face,which is of great significance for further improving the level of intelligence in fully mechanized caving mining.In this dissertation,there are 47 figures,24 tables and 97 references.