The Research Of Coal Gangue Identification And Separation Technique Based On Image Processing Technology

Coal resources occupy an important position in Chinese energy,and in the coal production,the direct exploitation of coal is called raw coal.It is inevitable to contain gangue in the raw coal.The selection of gangue in the coal preparation plant is separating the large gangue from the coal.It is an necessary procedure to separate the coal and gangue in all coal preparation plant.The present separation method of coal and gangue in the coal preparation plant is artificial separation.However in artificial separation,poor separation environment,workers labor intensity,low separation efficiency and so on are the main problems.With the development of machine vision technology,image processing technology has been used to study the separation of coal and gangue.In this paper,coal and gangue from one energy group plant was targeted for the research on coal and gangue identification and separation technology based on image processing.The main research contents are as follows:(1)The image of coal and gangue was pretreated based on image processing technology.The main processing included image grayscale,smoothing,sharpening and image segmentation.The weighted average method is used to process the image for image grayscal;the median filtering,adaptive median filtering and wavelet denoising are compared in the image smoothing,and the experimental result shows that the adaptive median filtering method is more effective;the Laplasse operator method was used to sharpen the image,and the adaptive thresholding method was for image segmentation.(2)The gray information and texture characteristics of coal and gangue image were studied in this paper in order to identify coal and gangue.The gray mean value,the gray variance,the smoothness,the three order moment and the consistent characteristic parameters of the coal and gangue image gray histogram were extracted based on the method of gray information;the energy,contrast,correlation and entropy feature parameters were extracted by using gray level co-occurrence matrix based on the method of texture feature.By analying the experimental results,the effective feature parameters were selected to form the feature vector,which was used for identifying coal and gangue.(3)The support vector machine was used to identify the coal and gangue.Firstly the principle of support vector machine was discussed,and then the parameters of SVM were optimized by particle swarm algorithm,finaly the optinal parameters of SVM was used to identify coal and gangue.The experimental results show that feature vector which was mixed with the gray information and texture feature can well descripe the features of coal and gangue.The classifier can well identify the coal and gangue by choosing the mixed feature vector as the input of support vector machine.(4)In this paper,the centroid method was used to locate the position of coal and gangue in the image.Firstly the adaptive threshold algorithm was used to binarise the gray image of coal and gangue and then the edge of the target region is extracted from the coal and gangue image.In edge detection,the differential gradient algorithm and the Canny algorithm was studied.By analising the experimental simulation results,the Roberts operator algorithm in differential gradient algorithm was choosed for the edge detection.Then,the dilation and erosion operations in the morphological knowledge were used to eliminate the connectivity of the holes and the connected regions in the target region.Finally,the centroid method was used to extract the centroid coordinates of the target area in the coal gangue image.(5)In this paper,the overall design of the coal and gangue identification and sorting system was studied.It includes hardware components and software communication of the system.The recognition and sorting technology research on coal and gangue based on image processing technology has a good effect in the laboratory.It is proved the correctness of the basic theory.It is important for achieving the automatic recognition and separation of coal and gangue.