Application Of Multispectral Imaging And Cluster Analysis In Coal And Gangue Recognition

As the most widely distributed fossil energy in the world,coal occupies an important position in the world’s energy.With the increasingly close connection and development of the coal industry and the information technology industry,the coal industry has begun to develop toward informatization,modernization,and intelligence.This research originates from the actual coal mine safety production,mainly relying on the sub-projects of the national key research and development plan,to explore the application of multi-spectral imaging combined with cluster analysis in coal gangue identification.This paper selects 200 samples of coal and gangue from a certain mining area in Huainan,introduces various methods of coal gangue sorting,and selects multi-spectral imaging technology as the means of collecting coal gangue images.Then the collected coal gangue image information is extracted by three algorithms of local binary mode,directional gradient histogram,and rotation invariant feature conversion fusion bag of words to extract the key features of the image information,and the rotation invariant feature conversion algorithm is selected.Feature information of coal gangue image.The data after feature extraction is compared with each band,and the coal gangue image with a wavelength of 954.412 mm is selected as the input data of the clustering algorithm.The processed coal gangue information is predicted by five clustering algorithms: K-means clustering,fuzzy C-means clustering,self-organizing competitive network,self-organizing feature map neural network,and spectral clustering,and different algorithms are obtained for coal The recognition accuracy rate of the multi-spectral image information of the gangue.Finally,the accuracy rate,recall rate and F1-Score are further compared with the three performance indicators,and the fuzzy C-means clustering algorithm is selected as the optimal algorithm of this research.The recognition accuracy rate is 96%,and the single run time is 0.016 s.Through the research of this article,the advanced multispectral technology is finally selected to collect multispectral images of coal gangue,and then the rotation invariant feature conversion fusion bag-of-words algorithm is used for feature extraction,and the 20 th wavelength with a wavelength of 954.412 mm is selected by comparing the accuracy of each band.Band analysis is performed,and finally the fuzzy C-means clustering algorithm is selected as the clustering algorithm in this paper.This paper presents for the first time the application of multi-spectral imaging technology and cluster analysis to coal gangue identification,which expands the development direction of coal gangue identification technology.The final result shows that the research is worthy of in-depth.Compared with other coal gangue identification methods,this study has a higher recognition accuracy and a short running time.What distinguishes it from other supervised learning algorithms is that cumbersome steps such as input tags are omitted.At the same time,the types of coal and gangue in different mining areas are different.The scalability of the clustering algorithm is better demonstrated.The application of clustering algorithm to coal gangue identification provides a new idea for the accurate identification of coal gangue,and it also enriches the application of clustering algorithms.Figure [55] table [11] reference [93]...