传统选煤常采用人工的方法，该方法效率低下且成本较高。机器视觉作为人工智能领域中的主要研究热点，近些年发展日趋成熟，在航空航天、人脸识别、无人驾驶等方面已得到广泛应用。本文将改进后的果蝇优化算法（Fruit Fly Optimization Algorithm，FOA）搜索SVM分类模型中的最优参数，用于解决煤与矸石的分类识别问题，为机器视觉技术在煤矸识别中的应用提供研究基础。本课题的主要研究工作如下：

针对传统分水岭图像分割算法在处理煤矸图像时效果不理想、存在过分割等问题，本文通过对煤矸图像目标区域和背景区域进行标记实现传统分水岭算法的优化，提出了一种基于距离标记的分水岭分割算法，将改进后的算法用于煤矸图像的分割，分割效果与传统的方法相比较分割效果较好，消弱了矸石表面附有煤粉、光照不良等情况的影响。

本文分别对煤和矸石的灰度特征与纹理特征进行了研究，并进行相关度分析，去掉部分不必要的特征，减小了数据量，增加了识别效率。通过对基于灰度特征、纹理特征以及灰度-纹理联合特征三种分类方法结果对比得出，基于灰度-纹理联合特征的分类方法效果最好。将SVM、BP以及KNN三种算法结合灰度-纹理联合特征对煤和矸石分类，经实验对比可知，SVM具有更高的识别率，其准确率为94.0%。因此，SVM在煤矸石识别中更具前景和优势。

针对FOA搜索半径为固定值、味道浓度公式恒为正值，影响了算法的收敛精度、速度以及全局搜索能力，本文从递减半径策略、改进味道浓度公式及随机机制策略三个方面对FOA进行改进优化，提出了一种改进的FOA，将改进的FOA、LFOA、FOA分别在5个常用的测试函数上面进行寻优计算，仿真结果显示，改进后的算法精度更高、收敛速度更快且稳定性更好。

针对SVM在煤矸分类中的不足，本文使用改进的FOA对SVM进行优化，并结合Heart数据集对算法分类性能进行评价。分别将FOA-SVM、PSO-SVM、LFOA-SVM以及本文所提算法用于解决煤矸的识别分类问题，利用准确率、精确率、召回率以及F₁-score四个指标对分类结果进行评价，结果表明，使用本文所提算法分类结果良好，平均准确率达到了96.30%，平均精确率达到了94.47%，平均召回率达到了98.13%，平均F₁-score达到了96.10%，通过对比实验验证了本文所提算法的有效性。

基于MATLAB2018软件平台对煤矸图像进行分类，通过对分类结果进行对比，验证了本课题所改进分类算法可靠有效，具有一定的应用价值。

Traditional coal preparation often uses manual methods, which are inefficient and costly. As a major research hot spot in the field of artificial intelligence, machine vision has become more mature in recent years and has been widely used in aerospace, face recognition, and unmanned driving. In this paper, the improved Fruit Fly Optimization Algorithm (FOA) is used to search the optimal parameters in the SVM model to solve the classification and identification of coal and gangue, and a research foundation for the application of machine vision technology in identification of coal and gangue is provided. The main research work of this subject is as follows:

(1) In view of the unsatisfactory effect of traditional watershed image segmentation algorithm in processing coal gangue images and the problem such as over-segmentation, the target area and background area of coal gangue images are marked to realize the optimization of the traditional watershed algorithm, and a watershed segmentation algorithm based on distance marker is proposed in this paper. The improved algorithm is applied to the segmentation of coal gangue images, and the segmentation effect is better than that of the traditional method, which weakens the influence of coal powder and poor illumination on the surface of gangue.

(2) In this paper, the features of gray-scale and texture of coal and gangue are studied respectively, and the correlation analysis is carried out to remove some unnecessary features, reduce the amount of data, and increase the recognition efficiency. By comparing the results of three classification methods based on gray-scale feature, texture feature and gray-texture joint feature, it is concluded that the classification method based on gray-texture joint feature is the best. Combining the three algorithms of SVM, BP and KNN with gray-texture joint features to classify coal and gangue, the comparison of experiments shows that SVM has a higher recognition rate, with an accuracy rate of 94.0%. Therefore, SVM has more prospects and advantages in coal gangue identification.

(3) Aiming at the search radius of FOA is a fixed value, and the flavor concentration formula is always positive, which affects the algorithm's convergence accuracy, speed and global search ability, the flavor concentration formula and the random mechanism strategy are used to improve and optimize FOA, and an improved FOA is proposed. The improved FOA, LFOA, and FOA are optimized and calculated on 5 commonly used test functions respectively. The simulation results show that, the improved algorithm has higher accuracy, faster convergence speed and better stability.

(4) Aiming at the shortcomings of SVM in identification of coal and gangue, the improved FOA was used in this paper to optimize SVM, and the classification performance of the algorithm was evaluated in combination with the HEART data set. FOA-SVM, PSO-SVM, LFOA-SVM and the algorithm proposed in this paper are used to solve the problem of identification of coal and gangue. The four indicators of accuracy, precision, recall and F1-score are used to evaluate the classification results. The results show that the classification results using the algorithm proposed in this article are good, with an average accuracy rate of 96.30%, an average precision rate of 94.47%, an average recall rate of 98.13%, and an average F1-score of 96.10%. This is verified by comparative experiments. The effectiveness of the algorithm proposed in this article.

The coal gangue images are classified based on the MATLAB2018 software platform. By comparing the classification results, it is verified that the improved classification algorithm in this subject is reliable and effective, and has certain application value.

[1]李珂. 我国煤炭资源清洁高效利用现状及对策建议[J]. 内蒙古煤炭经济, 2020(15): 175-176.

[2]谢和平, 吴立新, 郑德志. 2025年中国能源消费及煤炭需求预测[J]. 煤炭学报, 2019, 44(07): 1949-1960.

[3]高峰. 我国煤炭需求总量分析与预测[J]. 煤炭经济研究, 2014, 34(4): 10-13.

[4]Beata Jabłońska, Andriy V. Kityk, Mark Busch, Patrick Huber. The structural and surface properties of natural and modified coal gangue[J]. Journal of Environmental Management, 2017, (09): 80-90.

[5]吕东. 煤矸石淋溶液中的重金属对土壤环境的累积影响[J]. 山西化工, 2020, 40(03): 173-174+177.

[6]于海选, 廖祥国, 宋宝强, 郜二鹏. 智能化新技术在选煤厂的应用[J]. 煤炭加工与综合利用, 2021(02): 47-49.

[7]陈岩, 陈田佳, 陈光林, 李雷, 李志强. 基于射线透射技术的干法智能煤矸分选系统的应用[J]. 煤炭加工与综合利用, 2021(01): 60-62.

[8]权春锋, 马利云. 基于X射线的高含矸煤一体化干法选煤技术研究[J]. 矿山机械, 2020, 48(04): 73-76.

[9]P.Vithu, J. A. Moses. Machine vision system for food grain quality evaluation: A review[J], Trends in Food Science & Technology, 2016, 01: 13-20.

[10]徐志强, 吕子奇, 王卫东, 张康辉, 吕海梅. 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.

[11]Dong J, Lu B, He K, et al. Assessment of hatching properties for identifying multiple duck eggs on the hatching tray using machine vision technique[J]. Computers and Electronics in Agriculture, 2021, 184: 106076.

[12]兰俊斯. 机器视觉技术在汽车集装箱自助计量环节的应用[J]. 工业控制计算机, 2021, 34(02): 66-67+70.

[13]姚静姝. 计算机视觉在人脸识别领域中的应用[J]. 电子技术与软件工程, 2020(17): 128-129.

[14]Ganesh V N, Acharya S G, Bhat S, et al. Machine Vision Robot with Real Time Sensing[J]. Journal of Advancements in Robotics, 2020, 1(03): 30-34.

[15]曾仕峰, 吴锦均, 叶智文, 叶妙欣, 赖怡雯, 丁凡. 基于机器视觉的无人驾驶系统设计[J]. 电子世界, 2020(05): 197.

[16]Dun H, Tiberius C C J M, Janssen G J M. Positioning in a multipath channel using OFDM signals with carrier phase tracking[J]. IEEE Access, 2020, 08: 13011-13028.

[17]Udendhran R, Balamurugan M, Suresh A, et al. Enhancing image processing architecture using deep learning for embedded vision systems[J]. Microprocessors and Microsystems, 2020, 76: 103094.

[18]杨庆福. 国外光电选矿及其进展[J]. 化工矿物与加工, 1986(4): 50-54.

[19]丰建荣. 煤和矸石井下破碎分选理论及实验研究[D]. 太原: 太原理工大学, 2006.

[20]David M Hobson, Robert M Carter, Yong Yan. Differentiation between Coal and Stone through Image Analysis of Texture Features[J]. IEEE International Workshop on Imaging Systems and Techniques, 2007, 1-4.

[21]Singh V, Singh T N, Singh V. Image processing applications for customized mining and ore classification[J]. Arabian Journal of Geosciences, 2011, 4(7): 1163-1171.

[22]Claudio A. Perez, et al. Ore grade estimation by feature selection and voting using boundary detection in digital image analysis[J]. International Journal of Mineral Processing. 2011, 101(01): 28-36.

[23]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.

[24]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, 08: 55204-55220.

[25]Alfarzaeai M S, Niu Q, Zhao J, et al. Coal/Gangue Recognition Using Convolutional Neural Networks and Thermal Images[J]. IEEE Access, 2020, 8: 76780-76789.

[26]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.

[27]贾木斯. 基于X射线和光纤传感器的煤与矸石分选控制系统研究[D]. 沈阳: 东北大学, 2017.

[28]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(4): 27-31.

[29]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(2): 98-101.

[30]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.

[31]张万枝, 王增才. 基于视觉技术的煤岩特征分析与识别[J]. 煤炭技术, 2014, 33(10): 272-274.

[32]曹现刚, 薛祯也. 基于迁移学习的GoogLenet煤矸石图像识别[J]. 软件导刊, 2019, 18(12): 183-186.

[33]李曼, 段雍, 曹现刚等. 煤矸分选机器人图像识别方法和系统[J]. 煤炭学报, 2020, 45(10): 3636-3644.

[34]Li C Y, Guo J C, Cong R M, et al. Underwater Image Enhancement by Dehazing With Minimum Information Loss and Histogram Distribution Prior[J]. IEEE Transactions on Image Processing, 2016, 25(12): 5664-5677.

[35]Yao Z, Lai Z, Wang C. Image Enhancement Based on Equal Area Dualistic Sub-image and Non-parametric Modified Histogram Equalization Method[C]//2016 9th International Symposium on Computational Intelligence and Design (ISCID). IEEE, 2016, 1: 447-450.

[36]Maki S, Konno T, Maeda H. Image enhancement in computerized tomography for sensitive diagnosis of liver cancer and semiquantitation of tumor selective drug targeting with oily contrast medium[J]. Cancer, 2015, 56(4): 751-757.

[37]Gu Q, Lü S, Jiang M, et al. Phase error correction method based on the Gaussian filtering algorithm and intensity variance[J]. Optoelectronics Letters, 2021, 17(4): 221-225.

[38]Hai X, Cao S, Cui S, et al. Image Filter Processing Algorithm Analysis and Comparison[C]//Journal of Physics: Conference Series. IOP Publishing, 2021, 1820(1): 012192.

[39]Yu Feng, Gong Xinhui, Li Hanpeng, et al. Differential cryptanalysis of image cipher using block-based scrambling and image filtering[J]. Information Sciences, 2021, 554: 145-156.

[40]Sultana F, Sufian A, Dutta P. Evolution of image segmentation using deep convolutional neural network: a survey[J]. Knowledge-Based Systems, 2020, 201: 106062.

[41]LaLonde R, Xu Z, Irmakci I, et al. Capsules for biomedical image segmentation[J]. Medical image analysis, 2021, 68: 101889.

[42]Yu H, He F, Pan Y. A scalable region-based level set method using adaptive bilateral filter for noisy image segmentation[J]. Multimedia Tools and Applications, 2020, 79(9): 5743-5765.

[43]Bai M, Urtasun R. Deep watershed transform for instance segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2017: 5221-5229.

[44]雷涛, 张肖, 加小红, 等. 基于模糊聚类的图像分割研究进展[J]. 电子学报, 2019, 47(8): 1776-1791.

[45]张海涛, 李雅男. 阈值标记的分水岭彩色图像分割[J]. 中国图象图形学报, 2015, 20(12): 1602-1611.

[46]Yang J, Kang Z, Cheng S, et al. An individual tree segmentation method based on watershed algorithm and three-dimensional spatial distribution analysis from airborne LiDAR point clouds[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2020, 13: 1055-1067.

[47]Aghmadi A, El Hani S, Mediouni H, et al. Hybrid Solar Forecasting Method Based on Empirical Mode Decomposition and Back Propagation Neural Network[C]//E3S Web of Conferences. EDP Sciences, 2021, 231: 02001.

[48]Santos W A, Vettorato M C, Fogaça J L, et al. B-mode ultrasonography and gray scale histogram for evaluation of the nuchal ligament in Quarter horse[J]. Pesquisa Veterinária Brasileira, 2019, 39(8): 663-667.

[49]陆兴华, 刘铭原, 龙庆佳, 等. 基于灰度直方图的运动目标特征检测算法[J]. 计算机与现代化, 2019, (6): 71-75.

[50]李劭杰, 王晨, 史崯. 基于多特征随机森林的恶意代码检测[J]. 计算机应用与软件, 2020, 37(10): 328-333.

[51]Venkatesvara Rao N, Venkatavara Prasad D, Sugumaran M. Real-time video object detection and classification using hybrid texture feature extraction[J]. International Journal of Computers and Applications, 2021, 43(2): 119-126.

[52]Juan Martinez, Jose D. Luis-Delgado, Luis M. Tomas-Balibrea, Automatic system for quality-based classification of marble textures[J], IEEE Transactions On Systems, Man, And Cybernetics-oart C: Applications And Reviews, 2005, 35(4): 488-497.

[53]Pan W T. A new fruit flfly optimization algorithm: taking the fifinancial distress model as an example[J]. Knowledge-Based Systems, 2012, 26: 69-74.

[54]王林, 吕盛祥, 曾宇容. 果蝇优化算法研究综述[J]. 控制与决策, 2017, 32(07): 1153-1162.

[55]韩虎. 果蝇优化算法的分析[J]. 计算机系统应用, 2017, 26(2): 9-17.

[56]张铸, 饶盛华, 张仕杰. 一种新型的混沌步长果蝇优化算法[J]. 计算机工程与科学, 2020, 42(04): 755-760.

[57]李宗岳, 陈志军, 李名远. 基于混沌扰动策略的果蝇优化算法[J]. 微电子学与计算机, 2016, 33(07): 64-68.

[58]孙瑶琴. 基于改进FOA优化的SVM在故障诊断中的应用[J]. 机械强度, 2017, 39(2): 285-290．

[59]许昊. 基于改进的果蝇算法优化BP神经网络在空气质量预测中的应用研究[D]. 江西: 南昌大学, 2020.

[60]吴小文, 擎. 果蝇算法和5种群智能算法的寻优性能研究[J]. 火力与指挥控制, 2013, 38(04): 17-20+25.

[61]刘立群, 韩俊英, 代永强, 等. 果蝇优化算法优化性能对比研究[J]. 计算机技术与发展, 2015, 25(8): 94-98.

[62]Jan Zizka, Frantisek Darena, Arnost Svoboda. Text Mining with Machine Learning: Principles and Techniques[M]. CRC Press: 2020-01-15.

[63]Göreke V, Sarı V, Kockanat S. A novel classifier architecture based on deep neural network for COVID-19 detection using laboratory findings[J]. Applied Soft Computing, 2021, 106: 107329.

[64]Hameed S S, Muralidharan V, Ane B K. Comparative analysis of fuzzy classifier and ANN with histogram features for defect detection and classification in planetary gearbox[J]. Applied Soft Computing, 2021, 106: 107306.