带式输送机是煤矿井下生产系统的主运输设备，采用机器视觉可以对带式输送机的带面矸石、杂物和皮带表面损伤进行有效监测，从而为机器人矸杂分拣和皮带安全管控提供基础。因此，本文从带式输送机的带面监测图像入手，对图像去尘雾、去运动模糊以及矸石与杂物的目标检测方法进行了深入研究，以实现运煤带面矸石及杂物的高精度智能检测。

       针对影响煤矿安全生产及运输的主要因素如环境和矸杂异物等以及所带来的危害进行了分析，从监测图像入手，进行了总体方案的设计，结合图像去尘雾、去运动模糊及矸杂智能检测方法和应用等方面的国内外研究现状，搭建图像数据采集平台，并根据不同影响因素以及特点构建不同的数据集，同时对构建的不同数据集进行全面的统计和分析，为带式输送机运煤带面图像增强及矸杂智能检测奠定基础。

       针对煤矿井下皮带运输系统受粉尘、水雾和低照度环境影响，所采集的图像因大气散射的作用而产生严重降质，带式输送机带面物体的特征难以辨认，视觉效果差的问题。围绕大气散射模型提出一种暗亮通道分割融合的图像去尘雾及增强方法，通过全局大气光强估计、暗亮通道透射率估计及分割融合、亮度增强及饱和度矫正以解决粉尘、水雾及低照度环境对图像视觉效果和质量的影响，利用主观视觉及客观指标如平均梯度和信息熵等并与多种去尘雾方法进行比较，结果均呈现较优，完成图像的去尘雾研究，为后续的目标检测提供高质量图像。

       针对煤矿井下相机与带式输送机上的目标物体相对运动而造成的图像产生拖影、目标物体轮廓不清晰及目标难以辨认的问题。以生成对抗网络DeblurGAN-v2为基础，通过设计三尺度鉴别器网络结构、将结构相似性损失以及梯度损失引入生成器损失函数的计算中，将WGAN-GP的鉴别器损失函数引入，解决目标物体的拖影问题，完成对原始网络模型的改进，利用主观视觉及客观指标如图像亮度及图像梯度标准差等并与多种去运动模糊方法进行比较，结果均呈现较优，完成图像的去运动模糊研究，为后续的目标检测提供高质量图像。

       针对带式输送机运煤带面矸杂异物智能监测和数据分析处理的需求及矸杂异物检测结果显示问题，选择YOLOv5作为目标检测的方法，并验证了图像处理及本文所提处理方法对于带式输送机带面矸杂异物目标检测功能增强的有效性，同时开发设计了基于PyQt5的图像处理-非煤异物检测系统软件，并选择Raspberry Pi 4B来作为软件系统的载体，对图像处理与检测前后的图像及检测信息进行对比和存储，为图像处理方法的改进以及矸杂异物的分拣工作奠定基础。

       The belt conveyor is the main transportation equipment of the underground production system in coal mine. The machine vision can effectively monitor the gangue, debris and belt surface damage of the belt conveyor, so as to provide the basis for robot-based gangues sorting and belt safety control. Therefore, this article starts with the surface monitoring image of the belt conveyor, and conducts the in-depth research on the dust fog removal, motion blur removal, and target detection methods of gangues and impurities image, in order to achieve the high-precision intelligent detection of gangues and impurities on the coal belt.

       The main factors affecting the safety production and transportation of coal mine, such as environment and gangues foreign bodies, and the harm caused by them are analyzed. Starting from the monitoring images, the overall scheme is designed. Combined with the research status of image dust removal, motion blurring removal and gangues intelligent detection methods and applications at home and abroad, an image data acquisition platform is built, and different image data sets are constructed according to different influencing factors and characteristics. At the same time, the comprehensive statistics and analysis of different data sets are carried out, which lays a foundation for image enhancement and gangues intelligent recognition of coal mine belt conveyor.

       In view of the influence of dust, water mist and low illumination environment on the underground belt transportation system of coal mine, the collected images are seriously degraded due to the effect of atmospheric scattering, and the characteristics of the belt surface objects of the belt conveyor are difficult to identify and the visual effect is poor. Based on the atmospheric scattering model, an image dust removal and enhancement method based on dark-bright channel segmentation and fusion is proposed. Through the global atmospheric light intensity estimation, dark-bright channel transmittance estimation and segmentation fusion, brightness enhancement and saturation correction, the influence of dust, water mist and low illumination environment on image visual effect and quality is solved. The subjective vision and objective indicators such as average gradient and information entropy are used and compared with various dust removal methods. The results are better, and the dust and fog removal research of the image is completed, which provides high-quality images for subsequent target detection.

        Aiming at the problems of image smear, unclear contour of target object and difficult detection of target caused by the relative high-speed movement of target object on camera and belt conveyor in underground coal mine. Based on the generative adversarial network DeblurGAN-v2, the three-scale discriminator network structure is designed, and the structural similarity loss and gradient loss are introduced into the calculation of the generator loss function. The discriminator loss function of WGAN-GP is introduced to solve the smear problem of the target object, and the improvement of the original network model is completed. The subjective vision and objective indicators such as image brightness and image gradient standard deviation are used and compared with a variety of motion deblurring methods. The results are better, and the motion deblurring research of the image is completed, which provides high-quality images for subsequent target detection.

       Aiming at the requirements of intelligent monitoring and data analysis and processing of foreign object in coal belt of belt conveyor and the problem of detection results of foreign object in coal belt, YOLOv5 is selected as the method of target detection, and the effectiveness of image processing and the processing method proposed in this article for the enhancement of target detection function of foreign object in coal belt of belt conveyor is verified. At the same time, an image processing-non-coal foreign object detection system software based on PyQt5 is developed and designed, and Raspberry Pi 4B is selected as the carrier of the software system. The image and detection information before and after image processing and detection are compared and stored. It lays a foundation for the improvement of image processing methods and the sorting of foreign objects.

[1] 刘峰, 曹文君, 张建明, 等. 我国煤炭工业科技创新进展及“十四五”发展方向[J]. 煤炭学报, 2021,46(01):1-15.

[2] 王国法, 杜毅博. 智慧煤矿与智能化开采技术的发展方向[J]. 煤炭科学技术, 2019,47(01):1-10.

[3] 刘峰, 郭林峰, 赵路正, 等. 双碳背景下煤炭安全区间与绿色低碳技术路径[J]. 煤炭学报, 2022,47(01):1-15.

[4] 葛世荣, 郝尚清, 张世洪, 等. 我国智能化采煤技术现状及待突破关键技术[J]. 煤炭科学技术, 2020,48(07):28-46.

[5] 王国法, 范京道, 徐亚军, 等. 煤炭智能化开采关键技术创新进展与展望[J]. 工矿自动化, 2018,44(2):5-12.

[6] “十四五”能源领域科技创新规划[J]. 中国科技奖励, 2022,(09):23-42.

[7] 曹亦俊, 刘敏, 邢耀文, 等. 煤矿井下选煤技术现状和展望[J]. 采矿与安全工程学报, 2020,37(1):192-201.

[8] 邓洪刚, 王晓薇. 多煤种混配条件下除杂物系统的应用与改进[J]. 中国新技术新产品, 2011(7):32-32.

[9] 李文利, 朱子祺. 石圪台选煤厂杂物治理措施[J]. 煤炭加工与综合利用, 2011(4):9-10.

[10] 张万青, 李晓燕. 完善除杂设施严把产品质量关[J]. 煤质技术, 2009(2):31-32.

[11] 薛峰, 刘则庆, 王晋才, 等. 图像识别+机器人除杂系统在涡北选煤厂的应用[J]. 煤炭加工与综合利用, 2020(12):16-18.

[12] 毛善君, 崔建军, 王世斌, 等. 煤矿智能开采信息共享管理平台构建研究[J]. 煤炭学报, 2020,45(06):1937-1948.

[13] 程德强, 徐进洋, 寇旗旗, 等. 融合残差信息轻量级网络的运煤皮带异物分类[J]. 煤炭学报, 2022,47(03):1361-1369.

[14] 许永祥, 李申龙, 王国法, 等. 特厚坚硬煤层超大采高综放首采工作面智能化技术[J]. 煤炭科学技术, 2020,48(7):186-194.

[15] 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(4):7-12.

[16] 刘洋, 张杰, 张慧, 等. 一种改进的Retinex算法在图像去雾中的研究与应用[J]. 计算机科学, 2018,45(S1):242-243+251.

[17] Kumar A, Rajib K J, Nishchal N K. An improved Gamma correction model for image dehazing in a multi-exposure fusion framework[J]. Journal of Visual Communication and Image Representation, 2021:103122-103136.

[18] Yu Z, Sun B Y, Liu D, et al. STRASS Dehazing:Spatio-Temporal Retinex-Inspired Dehazing by an Averaging of Stochastic Samples[J]. Journal of Renewable Materials, 2022,10(5):1381-1395.

[19] 智宁, 毛善君, 李梅, 等. 基于深度融合网络的煤矿图像尘雾清晰化算法[J]. 煤炭学报, 2019,44(02):655-666.

[20] Qian W, Zhou C, Zhang D Y. FAOD-Net: A Fast AOD-Net for Dehazing Single Image[J]. Mathematical Problems in Engineering, 2020:1-11.

[21] Yan B N, Yang Z Z, Sun H Z, et al. ADE-CycleGAN: A Detail Enhanced Image Dehazing CycleGAN Network[J]. Sensors, 2023,23(6):3294-3294.

[22] Land E H. The retinex theory of color vision[J]. Scientific American, 1978,237(6):108-128.

[23] McCartney E J, Freeman F H. Optics of the Atmosphere: Scattering by Molecules and Particles[J]. Physics Today, 1977,30(5):76-77.

[24] Narasimhan S G, Nayar S K. Contrast restoration of weather degraded images[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2003,25(6):713-724.

[25] Narasimhan S G, Nayar S K. Vision and the Atmosphere[J]. International Journal of Computer Vision, 2002,48(3):233-254.

[26] He K M, Sun J, Tang X O. Single Image Haze Removal Using Dark Channel Prior[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2011,33(12):2341-2353.

[27] He K M, Sun J, Tang X O. Guided Image Filtering[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013(6):1397-1409.

[28] Tan R T. Visibility in bad weather from a single image[C]//2008 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2008), 24-26 June 2008, Anchorage, Alaska, USA. IEEE, 2008.

[29] Fattal R. Single image dehazing[J]. ACM Transactions on Graphics (TOG), 2008,27(3):1-9.

[30] Zhou H, Zhang Z Z, Liu Y, et al. Single Image Dehazing Algorithm Based on Modified Dark Channel Prior: Regular Section[J]. IEICE Transactions on Information and Systems, 2021, E104D(10). 1758-1761.

[31] Chen Z X, Ou B N, Tian Q Y. An improved dark channel prior image defogging algorithm based on wavelength compensation[J]. Earth Science Informatics, 2019,12(4):501-512.

[32] Zhu Q S, Mai J M, Shao L. A Fast Single Image Haze Removal Algorithm Using Color Attenuation Prior[J]. IEEE transactions on image processing: a publication of the IEEE Signal Processing Society, 2015,24(11):3522-3533.

[33] Yu T, Song K, Miao P, et al. Nighttime Single Image Dehazing via Pixel-Wise Alpha Blending[J]. IEEE Access, 2019,7:114619-114630.

[34] 赵猛, 任志浩, 褚海峰, 等. 基于大气散射模型的采煤工作面尘雾图像清晰化算法[J]. 煤炭学报, 2023,48(08):3312-3322.

[35] Lalonde J F, Efros A A, Narasimhan S G. Estimating the Natural Illumination Conditions from a Single Outdoor Image[J]. International Journal of Computer Vision, 2012,98(2):123-145.

[36] 赵波, 李洪平, 金汝宁, 等. 一种基于自适应大气光和加权引导滤波的夜间图像去雾算法[J]. 工程科学与技术, 2023,55(03):225-234.

[37] 陈彬, 谭新成. 加权聚合图像引导滤波[J]. 计算机工程与设计, 2020, 41(07):1993-1997.

[38] Hansen P C. Rank-Deficient and Discrete Ill-Posed Problems:Numerical Aspects of Linear Inversion[M]. Philadelphia:SIAM,1997.

[39] Neelamani R, Choi H, Baraniuk R. ForWaRD: Fourier-wavelet regularized deconvolution for ill-conditioned systems[J]. IEEE Transactions on signal processing, 2004,52(2):418-433.

[40] Singh M K, Tiwary U S, Kim Y H. An adaptively accelerated Lucy-Richardson method for image deblurring[J]. EURASIP Journal on Advances in Signal Processing, 2008:1-10.

[41] Wen Y W, Ng M K, Ching W K. Iterative algorithms based on decoupling of deblurring and denoising for image restoration[J]. SIAM Journal on Scientific Computing, 2008,30(5):2655-2674.

[42] Goodfellow I , Pouget-Abadie J , Mirza M ,et al.Generative Adversarial Nets[J].MIT Press, 2014,27.

[43] Chakrabarti A. A neural approach to blind motion deblurring[C]//Computer Vision-ECCV 2016:14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part III 14. Springer International Publishing, 2016:221-235.

[44] Sun J, Cao W F, Xu Z B, et al. Learning a convolutional neural network for non-uniform motion blur removal[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2015:769-777.

[45] Nah S, Hyun Kim T, Mu Lee K. Deep multi-scale convolutional neural network for dynamic scene deblurring[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2017:3883-3891.

[46] Tao X, Gao H Y, Shen X Y, et al. Scale-recurrent network for deep image deblurring[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2018:8174-8182.

[47] Kupyn O, Budzan V, Mykhailych M, et al. Deblurgan: Blind motion deblurring using conditional adversarial networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2018:8183-8192.

[48] Kupyn O, Martyniuk T, Wu J, et al. Deblurgan-v2: Deblurring (orders-of-magnitude) faster and better[C]//Proceedings of the IEEE/CVF international conference on computer vision. 2019:8878-8887.

[49] Lin T Y, Dollár P, Girshick R, et al. Feature pyramid networks for object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2017:2117-2125.

[50] Liu Y B, Haridevan A, Schofield H, et al. Application of Ghost-DeblurGAN to Fiducial Marker Detection[C]//2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2022:6827-6832.

[51] Han K, Wang Y H, Tian Q, et al. GhostNet: More features from cheap operations[J]. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2020,1577-1586.

[52] 赵辉, 黄镖, 王红君, 等. 基于改进DeblurGAN-v2的运动模糊农作物害虫图像复原方法[J/OL]. 光电子·激光, 1-8.http://kns.cnki.net/kcms/detail/12.1182.O4.20240322.1024.005.html.

[53] Hou Q B, Zhou D Q, Feng J S. Coordinate attention for efficient mobile network design[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2021:13713-13722.

[54] 申斌学, 朱磊, 古文哲, 等. 煤矿智能干选系统设计方法研究[J]. 煤炭工程, 2021,53(5):17-22.

[55] 刘富强, 钱建生, 王新红, 等. 基于图像处理与识别技术的煤矿矸石自动分选[J]. 煤炭学报, 2000,25(5):534-537.

[56] 臧洁. 小波分析在煤矸石图像识别ARM分选控制系统中的应用[D]. 西安:西安科技大学, 2008.

[57] 孙继平, 佘杰. 基于小波的煤岩图像特征抽取与识别[J]. 煤炭学报, 2013,38(10):1900-1904.

[58] 孙继平, 杨坤. 一种煤岩图像特征提取与识别方法[J]. 工矿自动化, 2017,43(05):1-5.

[59] 郜亚松, 张步勤, 郎利影, 等. 基于深度学习的煤矸石识别技术与实现[J]. 煤炭科学技术, 2021,49(12):202-208.

[60] 王婉婷, 姜国龙, 褚云飞, 等. 从RCNN到YOLO系列的物体检测系统综述[J]. 齐鲁工业大学学报, 2021,35(5):9-16．

[61] Girshick R. Fast R-CNN[C]//Proceedings of the IEEE international conference on computer vision. 2015,7:1440-1448.

[62] Ren S, He K, Girshich R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017,39(6):1137-1149.

[63] 吕志强.复杂环境下煤矿皮带运输异物图像识别研究[D]. 徐州:中国矿业大学, 2020.

[64] 吴守鹏, 丁恩杰, 俞啸, 等. 基于改进FPN的输送带异物识别方法[J]. 煤矿安全, 2019,50(12):127-130.

[65] 李莹. 基于深度学习的煤矸石目标检测方法研究[D]. 西安:西安科技大学, 2020.

[66] Wei L, Anguelov D, Erhan D, et al. SSD: single shot multibox detector[C]. Proceedings of the 14th European Confer ence. Netherlands: Springer, 2016:21-37.

[67] Redmon J, Divvala S, Girshick R, et al. You only look once: unified, real-time object detection[C]. Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recog nition (CVPR). Las Vegas: IEEE, 2016:779-788.

[68] Wang Y B, Wang Y J, Dang L F. Video detection of foreign objects on the surface of belt conveyor underground coal mine based on improved SSD[J]. Journal of Ambient Intelligence and Humanized Computing, 2020:1-10.

[69] Li D Y, Wang G F, Zhang Y, et al. Coal gangue detection and recognition algorithm based on deformable convolution YOLOv3[J]. IET Image Processing, 2022,16(1):134-144.

[70] Redmon J, Farhadi A. YOLOv3:An Incremental Improvement[J]. Computer Vision and Pattern Recognition, 2018,276(7):126-134．

[71] 郝帅, 张旭, 马旭, 等. 基于CBAM-YOLOv5的煤矿输送带异物检测[J]. 煤炭学报, 2022,47(11):4147-4156.

[72] Woo S, Park J, Lee J Y, et al. Cbam:convolutional blockattention module[C]//European Conference on ComputerVision (ECCV), 2018:3-19.

[73] 毛清华, 李世坤, 胡鑫, 等. 基于改进YOLOv7的煤矿带式输送机异物识别[J]. 工矿自动化, 2022,48(12):26-32.

[74] Wang C Y, Bochkovskiy A, Liao H Y M. YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2023:7464-7475.

[75] Zeng Q L, Zhou G Y, Wan L R, et al. Detection of Coal and Gangue Based on Improved YOLOv8[J]. Sensors, 2024,24(4):1246.

[76] 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.

[77] 王家臣, 李良晖, 杨胜利, 等. 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.

[78] 任敬. 基于GAN的车牌图像去运动模糊方法研究[D]. 西安:陕西师范大学, 2021.

[79] Choi L K, You J, Bovik A C. Referenceless Prediction of Perceptual Fog Density and Perceptual Image Defogging[J]. IEEE transactions on image processing: a publication of the IEEE Signal Processing Society, 2015, 24(11):3888-3901.

[80] Wang X C, Sun X D, Song R X. Single Image Dehazing with V-transform and Dark Channel Prior[J]. Journal of Systems Science and Information, 2020,8(02):185-194.

[81] 牛宏侠, 王春智. 基于HSI空间的沙尘图像增强算法[J]. 北京交通大学学报, 2022,46(05):1-8.

[82] Wang S H, Luo G. Naturalness Preserved Image Enhancement Using a priori Multi-Layer Lightness Statistics.[J]. IEEE transactions on image processing: a publication of the IEEE Signal Processing Society, 2018,27(2):938-948.

[83] 陈文艺, 赵晓光. 一种自适应控制饱和度的色调映射方法[J]. 西安邮电大学学报, 2016,21(03):84-89.

[84] Sandler M, Howard A, Zhu M, et al. MobilenetV2: Inverted Residuals and Linear Bottlenecks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2018:4510-4520.

[85] Li C, Wand M. Precomputed real-time texture synthesis with markovian generative adversarial networks[C]//Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part III 14. Springer International Publishing, 2016:702-716.

[86] Simonyan K , Zisserman A .Very Deep Convolutional Networks for Large-Scale Image Recognition[J].Computer Science, 2014.DOI:10.48550/arXiv.1409.1556.

[87] 王旭晴, 魏伟波, 杨光宇, 等. 基于算法展开的图像盲去模糊深度学习网络[J].山东大学学报(工学版), 2023,53(06):35-46.

[88] Qi Q, Guo J C, Jin W P. Attention Network for Non-Uniform Deblurring[J].IEEE Access, 2020,8:100044-100057.

[89] Mao X D, Li Q, Xie H R, et al. Least Squares Generative Adversarial Networks[J].IEEE International Conference On Computer Vision, 2017:2794-2802.

[90] Gulrajani I, Ahmed F, Arjovsky M, et al. Improved Training of Wasserstein GANs[J]. Advances In Neural Information Processing Systems, 2017.DOI:10.48550/arXiv.1704.00028.

[91] Wang C Y, Yeh I H, Liao H Y. YOLOv9:Learning What You Want to Learn Using Programmable Gradient Information[J]. 2024.DOI:10.48550/arXiv.2402.13616.

[92] Bochkovskiy A, Wang Chienyao, Liao Hongyuan. YOLOv4:Optimal Speed and Accuracy of Object Detection[J]. 2020.DOI:10.48550/arXiv.2004.10934.

[93] Li C Y, Li L L, Jiang H L, et al. YOLOv6:A single-stage object detection framework for industrial applications[J]. 2022.DOI:10.48550/arXiv.2209.02976.