煤炭资源是我国重要的基础能源，采煤和洗煤的工艺过程中煤矸石含量的高低极大的影响了煤的纯度和质量，因此煤矸石的分选对提高煤炭自动化生产效率有着重要的意义。传统的煤矸分选方法效率低下、污染环境、成本过高，已经不能满足当今智慧矿山的发展需求。基于此，本文对传统分选方法进行总结和分析，从图像识别角度，提出采用图像纹理特征参数和深度学习技术进行煤矸石分选的方法，主要研究内容如下：
       论文以陕西韩城象山矿井所采集的煤炭及煤矸石为分析对象，围绕煤炭及煤矸石图像，开展对计算机智能辨识方法的研究。研究了对煤矸石图像常用的分割方法和传统纹理特征参数提取算法。通过对煤矸石图像进行同态滤波处理，增强煤和矸石的对比度；随后通过颜色空间转换得到HSV色彩，并利用K-means++聚类算法在HSV颜色空间对其进行图像分割，从而获得煤和矸石目标图像；在此基础上从图像纹理特征的角度对煤矸石图像进行纹理特征参数分析，分别提取了灰度共生矩阵的纹理特征参数和Tamura纹理特征参数，并结合BP神经网络以及支持向量机实现了煤矸石图像的分类识别。针对传统机器学习在矸石图像分选中的局限性，研究基于卷积神经网络的煤矸石分类模型。构建一种针对煤和煤矸石自动识别的深层卷积神经网络结构，通过自动提取特征参数，实现煤矸石图像自动分类识别。
       实验表明构建的特征参数结合BP神经网络以及支持向量机其识别准确率分别达到了76.2%与81.1%。构建的卷积神经网络模型的分类准确率可达到93.4%，对煤矸石和煤能够较好的进行分类识别，实现了无特征参数的煤矸石自动跟类识别，相比于传统的神经网络及支持向量机而言，本文所构建的卷积神经网络结构有效的提高了煤矸石识别的准确率。

~Coal resources are the important basic energy in China. The coal gangue content greatly affects the purity and quality of coal in the process of coal mining and coal washing. Therefore, the separation of coal gangue is of great significance to improve the efficiency of coal automation. The traditional coal gangue separation method is inefficient, polluting the environment and high cost, which can not meet the development needs of today's smart mine. Based on this, this paper summarizes and analyzes the traditional separation methods, and from the perspective of image recognition, puts forward the method of using image texture feature parameters and deep learning technology for coal gangue separation：
This paper takes coal and coal gangue collected from Xiangshan Coal Mine in Hancheng, Shaanxi Province as the analysis object, and studies the computer intelligent identification method around the image of coal and coal gangue.The commonly used segmentation methods and traditional texture feature extraction algorithms for gangue image are studied.By homomorphic filtering on the image of coal gangue, the contrast between coal and gangue is enhanced.Then HSV color is obtained by color space conversion, and the image is segmented in HSV color space using K-means++ clustering algorithm to get the target image of coal and gangue.Based on this, the texture feature parameters of coal gangue image are analyzed from the angle of image texture feature. The gray symbiosis matrix texture feature parameters and Tamura texture feature parameters are extracted, and the classification and recognition of coal gangue image is achieved by combining BP network and support vector machine.In view of the limitations of traditional machine learning in gangue image sorting, a gangue classification model based on convolution neural network is studied.A deep convolution neural network structure for automatic recognition of coal and gangue is constructed, and automatic classification and recognition of gangue image is achieved by automatically extracting feature parameters.
Experiments show that the recognition accuracy of the feature parameters combined with BP neural network and support vector machine is 76.2% and 81.1% respectively. The classification accuracy of the constructed convolution neural network model can reach 93.4%, which can better classify and recognize coal gangue and coal, and realize the automatic classification recognition of coal gangue without characteristic parameters. Compared with the traditional neural network and support vector machine, the convolution neural network structure constructed in this paper effectively improves the accuracy of coal gangue recognition.

[1]赵开功,李彦平.我国煤炭资源安全现状分析及发展研究[J]. 煤炭工程, 2018, 10: 185-189.

[2]王晴,刘锁,王继博,等.煤矸石粗集料混凝土抗氯离子渗透性能的研究[J].混凝土,2016(08):36-38.

[3]曹现刚,李莹,王鹏,等.煤矸石识别方法研究现状与展望[J].工矿自动化,2020,46(01):38-43.

[4]张凌智,代伟,马小平.重介质选煤过程先进控制：研究进展及展望[J].工矿自动化,2020,46(08):21-27+69.

[5]许铁卫,原野,王纪成.重介质选煤厂介耗控制途径探讨[J].煤炭工程,2019,51(02):66-69.

[6]孙娜, 王宏, 郭德. 重力选煤中分配曲线预测模型及应用[J]. 煤炭工程, 2017, 49(3): 63-66.

[7]王志刚,庞卫东,常建勇,等.新型大处理能力双侧进风跳汰分选技术[J].采矿与安全工程学报,2020,37(01):202-206.

[8]杨慧刚,乔志敏.基于X射线和机器视觉的煤与矸石分选系统设计[J].工矿自动化,2017,43(03):85-89.

[9]康利,黄金辉,刘畅.GDRT型γ射线智能干法分选系统在六家煤矿的应用[J].煤炭加工与综合利用,2017(03):22-24.

[10]王锐,桂志国,刘祎,等.基于X射线和结构光相机的煤矸石分拣方法研究[J/OL].中北大学学报(自然科学版),2021(02):123-128+134.

[11]王水生.基于振动特性分析的采煤机煤岩识别控制系统[J].工矿自动化,2015,41(05):83-87.

[12]张新,王正祥,武晨光,等.煤岩界面识别技术研究[J].煤矿机械,2015,36(02):3-5.

[13]王昕,胡克想,俞啸,等.基于太赫兹时域光谱技术的煤岩界面识别[J].工矿自动化,2017,43(01):29-34.

[14]葛世荣.采煤机技术发展历程(六)——煤岩界面探测[J].中国煤炭,2020,46(11):10-24.

[15]Reddy K G R, Tripathy D P. Separation of gangue from coal based on histogram thresholding[J]. International Journal of Technology Enhancements and Emerging Engineering Research, 2013, 1(4): 31-34.

[16]张世杰,陈泽华.基于梯度下降算法的煤矸石分割识别方法[J].煤矿机械,2014,35(08):128-130.

[17]丁泽海,薛斌,窦东阳.图像处理在煤矸石分选系统中的应用[J].煤矿机械,2017,38(03):173-175.

[18]胡倍倍,吕浩杰.基于改进直方图均衡化和SSR算法的灰度图像增强研究[J].量子电子学报,2017,34(03):286-292.

[19]武国平,梁兴国,胡金良,等. 图像处理和SVM应用于煤矸石分选的实验研究[J]. 信息技术,2019(1):97-102,107.

[20]程健,王东伟,杨凌凯,等.一种改进的高斯混合模型煤矸石视频检测方法[J].中南大学学报(自然科学版),2018,49(01):118-123.

[21]黄韶杰,刘建功.基于高斯混合聚类的煤岩识别技术研究[J].煤炭学报,2015,40(S2):576-582.

[22]于腾飞.基于相位一致性的煤矸边缘检测方法[J].电子器件,2016,39(06):1349-1352.

[23]吴开兴,宋剑.基于灰度共生矩阵的煤与矸石自动识别研究[J].煤炭工程,2016,48(02):98-101.

[24]吴开兴,宋剑.基于灰度和纹理的煤与矸石特征提取研究[J].煤炭技术,2015,34(11):305-307.

[25]陈雪梅,张晞,徐莉莉,等.煤与矸石分形维数的差异研究[J].煤炭科学技术,2017,45(07):196-199.

[26]谭春超,杨洁明.煤与矸石图像灰度信息和纹理特征的提取研究[J].工矿自动化,2017,43(04):27-31.

[27]王家臣,李良晖,杨胜利.不同照度下煤矸图像灰度及纹理特征提取的实验研究[J].煤炭学报,2018,43(11):3051-3061.

[28]陈立,杜文华,曾志强,等.基于小波变换的煤矸石自动分选方法[J].工矿自动化,2018,44(12):60-64.

[29]孙继平,陈浜.基于双树复小波域统计建模的煤岩识别方法[J].煤炭学报,2016,41(07):1847-1858.

[30]章华,李振璧,姜媛媛.基于图像纹理的煤岩识别研究[J].煤炭技术,2015,34(07):120-121.

[31]Hou W. Identification of coal and gangue by feed-forward neural network based on data analysis[J]. International Journal of Coal Preparation and Utilization, 2019, 39(1): 33-43.

[32]Eshaq R M A, Hu E, Li M, et al. Separation between coal and gangue based on infrared radiation and visual extraction of the YCbCr color space[J]. IEEE Access, 2020, 8: 55204-55220.

[33]Tripathy D P, Reddy K G R. Novel methods for separation of gangue from limestone and coal using multispectral and joint color-texture features[J]. Journal of The Institution of Engineers (India): Series D, 2017, 98(1): 109-117.

[34]伍云霞,田一民.基于字典学习的煤岩图像特征提取与识别方法[J].煤炭学报,2016,41(12):3190-319.

[35]Heaton J. Ian goodfellow, yoshua bengio, and aaron courville: Deep learning[J]. Genetic Programming and Evolvable Machines, 2017:1-3.

[36]Nasser M, Salim N, Hamza H, et al. Improved Deep Learning Based Method for Molecular Similarity Searching Using Stack of Deep Belief Networks[J]. Molecules, 2021, 26(1): 128.

[37]Hinton G E , Srivastava N , Krizhevsky A , et al. Improving neural networks by preventing co-adaptation of feature detectors[J]. Computer Science, 2012, 3(4):págs. 212-223.

[38]Szegedy C, Liu W, Jia Y, et al. Going deeper with convolutions[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2015: 1-9.

[39]He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2016: 770-778.

[40]Badrinarayanan V, Kendall A, Cipolla R. Segnet: A deep convolutional encoder-decoder architecture for image segmentation[J]. IEEE transactions on pattern analysis and machine intelligence, 2017, 39(12): 2481-2495.

[41]Boloor A, Garimella K, He X, et al. Attacking vision-based perception in end-to-end autonomous driving models[J]. Journal of Systems Architecture, 2020, 110: 101766.

[42]Getty N, Brettin T, Jin D, et al. Deep medical image analysis with representation learning and neuromorphic computing[J]. Interface Focus, 2021, 11(1): 20190122.

[43]Holmberg O G, Köhler N D, Martins T, et al. Self-supervised retinal thickness prediction enables deep learning from unlabelled data to boost classification of diabetic retinopathy[J]. Nature Machine Intelligence, 2020, 2(11): 719-726.

[44]Rajpurkar P, Park A, Irvin J, et al. AppendiXNet: Deep learning for diagnosis of appendicitis from A small dataset of CT exams using video pretraining[J]. Scientific reports, 2020, 10(1): 1-7.

[45]Azarang A, Kehtarnavaz N. Image fusion in remote sensing by multi-objective deep learning[J]. International Journal of Remote Sensing, 2020, 41(24): 9507-9524.

[46]Petrovska B, Zdravevski E, Lameski P, et al. Deep learning for feature extraction in remote sensing: A case-study of aerial scene classification[J]. Sensors, 2020, 20(14): 3906.

[47]Bhosle K, Musande V. Evaluation of deep learning CNN model for land use land cover classification and crop identification using hyperspectral remote sensing images[J]. Journal of the Indian Society of Remote Sensing, 2019, 47(11): 1949-1958.

[48]饶中钰,吴景涛,李明.煤矸石图像分类方法[J].工矿自动化,2020,46(03):69-73.

[49]王冠军,苏婷婷,刘文博,等.基于EAIDK的智能煤矸分拣系统设计[J].工矿自动化,2020,46(01):105-108.

[50]肖冬,毛亚纯,宋亮, 等.可见,近红外光谱和深度学习CNN-ELM算法的煤炭分类[J]. 光谱学与光谱分析, 2018, 38(7): 2107-2112.

[51]Pu Y, Apel D B, Szmigiel A, et al. Image recognition of coal and coal gangue using a convolutional neural network and transfer learning[J]. Energies, 2019, 12(9): 1735.

[52]赵明辉.一种煤矸石优化识别方法[J].工矿自动化,2020,46(07):113-116.

[53]曹现刚,费佳浩,王鹏,等.基于多机械臂协同的煤矸分拣方法研究[J].煤炭科学技术,2019,47(04):7-12.

[54]徐志强,吕子奇,王卫东,等.煤矸智能分选的机器视觉识别方法与优化[J].煤炭学报,2020,45(06):2207-2216.

[55]Hong H, Zheng L, Zhu J, et al. Automatic recognition of coal and gangue based on convolution neural network[J]. arXiv preprint arXiv:1712.00720, 2017.

[56]Xiao D, Le B T. Rapid analysis of coal characteristics based on deep learning and visible-infrared spectroscopy[J]. Microchemical Journal, 2020, 157: 104880.

[57]Fu Y, Aldrich C. Quantitative Ore Texture Analysis with Convolutional Neural Networks[J]. IFAC-PapersOnLine, 2019, 52(14): 99-104.

[58]蒋卫祥.增量式矿石自动化分拣系统研究[J].矿业研究与开发,2020,40(11):150-155.

[59]张斌,苏学贵,段振雄,等.YOLOv2在煤岩智能识别与定位中的应用研究[J].采矿与岩层控制工程学报,2020,2(02):94-101.

[60]宋庆军,李帅,黄兴成,等.基于卷积的放顶煤煤矸识别方法研究[J].煤炭技术,2021,40(04):25-27.