我国是世界煤炭生产大国，输煤皮带传输是煤炭开采中一个重要环节。由于开采条件和生产环节的特殊性，在实际煤炭运输过程中，输煤皮带在无煤或少煤情况下长时间高速运转会造成大量电能损耗，因此根据输煤量自动调控输煤皮带速率以实现煤矿智能化开采是一项重要的研究工作。本文提出了一种YOLOE-CA的输煤皮带煤量检测算法，并针对实际应用场景将算法进行嵌入式平台部署。

       本文采用多种优秀策略对输煤皮带煤量检测算法进行改进，有效地提升了算法性能。在主干网络部分，本文将Focus结构加入其中，采用以空间换时间的方式减少计算参数和模型参数量；并改进网络中的残差结构，增强主干网络提取特征的能力。在Neck网络部分，本文对PAN结构进行改进，使用BiFPN金字塔结构对不同尺寸的特征图进行加权特征融合，提高网络对重要特征的关注度；并在Neck网络前加入CBAM注意力机制，提升模型特征图的表达能力。在检测头部分，本文使用CA注意力机制代替ESE，并且用CIoU代替GIoU计算损失函数，提高检测精度。实验结果表明，与PP-YOLOE算法相比，改进算法煤量检测AP50提升了4.2%。

       本文采用自建数据集对模型进行训练，并通过TensorRT进行优化，将改进后的算法部署在NVIDIA Jetson TX2嵌入式平台。设计实现了输煤皮带自适应速度控制系统。 根据输煤皮带煤流量的大小，分为（0~15%）慢速、（15%~50%）中速、（大于50%）快速三个等级，从而分档次调整皮带的运行速度。该系统可根据输煤皮带中煤量的大小变化自动调节带速，节能效果明显，适用于煤矿生产需求。并且还在软件功能中加入紧急停止功能，应对紧急情况。

        China is a major coal producer in the world, and coal belt transmission is an important part of coal mining. Due to the particularity of mining conditions and production links, in the actual coal transportation process, the long-term high-speed operation of the coal conveyor belt in the absence of coal or less coal will cause a large amount of power loss. Therefore, it is an important research work to automatically adjust the speed of the coal conveyor belt according to the amount of coal transportation to realize the intelligent mining of coal mines. This paper proposes a YOLOE-CA coal belt coal quantity detection algorithm, and deploys the algorithm on an embedded platform for practical application scenarios.

        In this paper, a variety of excellent strategies are used to improve the coal quantity detection algorithm of coal conveyor belt, which effectively improves the performance of the algorithm. In the backbone network part, this paper adds the Focus structure to reduce the calculation parameters and model parameters by exchanging space for time. And improve the residual structure in the network to enhance the ability of the backbone network to extract features. In the Neck network part, this paper improves the PAN structure, and uses the BiFPN pyramid structure to perform weighted feature fusion on feature maps of different sizes to increase the network 's attention to important features. The CBAM attention mechanism is added before the Neck network to improve the expression ability of the model feature map. In the detection head part, this paper uses CA attention mechanism instead of ESE, and uses CIoU instead of GIoU to calculate the loss function to improve the detection accuracy. The experimental results show that compared with the PP-YOLOE algorithm, the AP _ 50 of the improved algorithm is increased by 4.2 %.

        In this paper, the self-built data set is used to train the model and optimized by TensorRT. The improved algorithm is deployed on the NVIDIA Jetson TX2 embedded platform. An adaptive speed control system for coal conveying belt is designed and implemented. According to the coal flow rate of the coal conveying belt, it is divided into three grades : ( 0 ~ 15 % ) slow speed, ( 15 % ~ 50 % ) medium speed and ( > 50 % ) fast speed, so as to adjust the running speed of the belt. The system can automatically adjust the belt speed according to the change of coal quantity in the coal conveyor belt, and the energy saving effect is obvious, which is suitable for coal mine production demand. And also add an emergency stop function to the software function to respond to emergencies.

[1] 朱立平, 蒋卫良. 适用于我国煤矿带式输送机典型机型的研究状[J]. 煤矿学报, 2010, (11): 1916-1920.

[2] 赵玉文. 带式输送机的现状和发展趋势[J]. 煤矿机械.

[3] 王福斌, 孙海洋, TU Paul. 边缘扩展的皮带撕裂支持向量机视觉检测[J]. 中国机械工程, 2019, 30(4):455-460.

[4] 宋伟刚, 王丹. 7.6km长距离带式输送机系统动力学的计算机仿真[J]. 煤炭学报, 2004, 29(2):249-253.

[5] 孙伟, 王慧, 杨海群. 带式输送机变频调速节能控制系统研究[J]. 工矿自动化, 2013, 39(4):98-101.

[6] 杨祥, 田慕琴, 李璐,等. 带式输送机输送带故障检测技术研究[J]. 煤矿机械, 2019, 40(2):133-136.

[7] 田原. 浅谈机器视觉技术在煤矿中的应用前景[J]. 工矿自动化, 2010, 36(5):30-32.

[8] 张铮, 王艳平, 薛桂香. 数字图像处理与机器视觉[M]. 北京: 人民邮电出版社, 2010:340-379.

[9] 杨小彬, 周世禄, 李娜,等. 深度学习及其在煤矿安全领域的应用[J]. 煤矿安全, 2019, 50(1):261-264.

[10] 丁秀明. 煤矿能耗调研分析和节能措施[J]. 煤炭技术, 2010, 29(3):243-245.

[11] 任中全, 王淼. 带式输送机节能调速控制系统设计[J]. 煤炭技术, 2016, 35(5):245-246.

[12] 孟凡芹, 王耀才. 煤矿井下带式输送机煤流图像识别方法的研究[J]. 煤炭学报, 2003, 28(1):23-24.

[13] 唐守锋, 史可, 仝光明,等. 一种矿井低照度图像增强算法[J]. 工矿自动化, 2021, 47(10):5.

[14] 赵永强, 饶元, 董世鹏,等. 深度学习目标检测方法综述[J]. 中国图象图形学报, 2020, 25(4):629-654.

[15] 代伟, 赵杰, 杨春雨,等. 基于双目视觉深度感知的带式输送机煤量检测方法[J]. 煤炭学报, 2017, 42(S2):547-555.

[16] 周俊宇, 赵艳明. 卷积神经网络在图像分类和目标检测应用综述[J]. 计算机工程与应用, 2017, 53(13): 34-41.

[17] Wei Y, Tian Q, Guo J, et al. Multi-vehicle detection algorithm through combining Harr and HOG features[J]. Mathematics and Computers in Simulation, 2019, 155:130-145.

[18] Chen S, Zhong S, Xue B, et al. Iterative scale-invariant feature transform for remote sensing image registration[J]. IEEE Transactions on Geoscience and Remote Sensing, 2020, 59(4):3244-3265.

[19] Zhao H, Zhan Z H, Lin Y, et al. Local binary pattern-based adaptive differential evolution for multimodal optimization problems[J]. IEEE transactions on cybernetics, 2019, 50(7):3343-3357.

[20] 杨笑, 王志章, 周子勇, 等. 基于参数优化 AdaBoost 算法的酸性火山岩岩性分类 [J]. 石油学报, 2019, 40(4):457-467.

[21] Seeja R D, Suresh A. Deep learning based skin lesion segmentation and classification of melanoma using support vector machine (SVM)[J]. Asian Pacific journal of cancer prevention: APJCP. 2019, 20(5):1555.

[22] 王森, 伍星, 张印辉,等. 基于深度学习的全卷积网络图像裂纹检测[J]. 计算机辅助设计与图形学学报, 2018, 30(5):115-12.

[23] Le Cun Y, Bottou L, Bengio Y, et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86(11):2278-2324.

[24] Krizhevsky A, Sutskever I, Hinton G E. Imagenet classification with deep convolutional neural networks[C]//Advances in neural information processing systems. 2012: 1097- 1105.

[25] Simonyan K, Zisserman A. Very Deep Convolutional Networks for Large-Scale Image Recognition[J], Computer Science. 2014:1-14.

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

[27] 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.

[28] Jie H, Li S, Gang S, et al. Squeeze-and-excitation networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020 42(8): 2011-2023.

[29] Iandola F N, Han S, Moskewicz M W, et al. SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and< 0.5 MB model size[J]. arXiv:1602.07360, 2016.

[30] Howard A G, Zhu M, Chen B, et al. Mobilenets: Efficient convolutional neural networks for mobile vision applications[J]. arXiv:1704.04861, 2017.

[31] Sandler M , Howard A , Zhu M , et al. MobileNetV2: Inverted Residuals and Linear Bottlenecks[C]. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2018:4510-4520.

[32] Howard A, Sandler M, Chu G, et al. Searching for mobilenetv3[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. IEEE, 2019:1314-1324.

[33] Zhang X, Zhou X, Lin M, et al. Shufflenet: An extremely efficient convolutional neural network for mobile devices[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2018:6848-6856.

[34] Han K, Wang Y H, Tian Q, et al. GhostNet: more features from cheap operations[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE, 2020:1580-1589.

[35] Redmon J, Farhadi A. YOLO9000: better, faster, stronger[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition(CVPR). IEEE, 2017:7263- 7271.

[36] Redmon J, Farhadi A. YOLOv3:An incremental improvement[J]. Optics Laser Technology, 2018, 23(03):44-57.

[37] Bochkovskiy A., Wang C. Y., Liao H Y M. YOLOv4: Optimal Speed and Accuracy of Object Detection [J]. arXiv: 2020, 2004:10934

[38] Liu W, Anguelov D, Erhan D, et al. Ssd: Single shot multibox detector[C]//European conference on computer vision. Springer, Cham, 2016: 21-37.

[39] Girshick R, Donahue J, Darrell T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. IEEE, 2014:580-587.

[40] Girshick R. Fast R-CNN[C]// IEEE International Conference on Computer Vision. IEEE Computer Society. IEEE, 2015:1440-1448.

[41] Ren S, He K, Girshick 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.

[42] Law H, Deng J. Cornernet: Detecting objects as paired keypoints[C]//Proceedings of the European conference on computer vision (ECCV). 2018: 734-750.

[43] Duan K, Bai S, Xie L, et al. Centernet: Keypoint triplets for object detection[C]//Proceedings of the IEEE/CVF international conference on computer vision. 2019: 6569-6578.

[44] Z. Tian, C. Shen, H. Chen, et al. FCOS: Fully convolutional one-stage object detection[C]// IEEE/CVF International Conference on Computer Vision (ICCV). IEEE, 2019:9627-9636.

[45] Zhang S, Chi C, Yao Y, et al. Bridging the gap between anchor-based and anchor-free detection via adaptive training sample selection[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2020: 9759-9768.

[46] Redmon J., Divvala S., Girshick R. et al. You Only Look Once: Unified, Real-Time Object Detection[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2016:779-788.

[47] Ge Z, Liu S, Wang F, et al. YOLOX: Exceeding YOLO Series in 2021[J]. arXiv preprint arXiv:2107.08430,2021.

[48] Xu S , Wang X , Lv W , et al. PP-YOLOE: An evolved version of YOLO[J]. arXiv:2203.16250 ,2022.

[49] Lin T, Goyal P, Girshick R, He K, Dollár P. Focal Loss for Dense Object Detection[C]. 2017 IEEE International Conference on Computer Vision(ICCV). IEEE,2017:2999- 3007.

[50] Everingham M, Eslami S M A, Gool L V, et al. The Pascal Visual Object Classes Challenge: A Retrospective[J]. International Journal of Computer Vision, 2015, 111(1):98-136.

[51] Lin T Y, Maire M, Belongie S, et al. Microsoft coco: Common objects in context[C] //European Conference on Computer Vision. Springer, Cham, 2014: 740-755.

[52] S. Yun, D. Han, S. Chun, et al. Cutmix: Regularization strategy to train strong classifiers with localizable features[C]// IEEE/CVF International Conference on Computer Vision (ICCV). IEEE, 2019:6022-603