| 360 | 0 | 162 |
| 下载次数 | 被引频次 | 阅读次数 |
传统的摇床选矿方法依赖工人凭借生产经验手动调整接矿板位置，这会导致人工成本升高且接矿结果不稳定.深度学习技术的发展为摇床的智能化应用提供了可能.YOLOv8目标检测模型以其高检测精度和鲁棒性在目标检测领域获得了广泛认可.然而，YOLOv8在处理复杂场景时存在局限性，如浅层特征提取不足和高阶特征图的通道表达能力不足等.为解决上述问题，提出浅层特征注意力模块(LLA)和深层特征注意力模块(HLA),分别增强了低阶和高阶特征图的表达能力，并通过引入C2f-faster模块降低模型参数量，实现轻量化.实验表明，本研究与五个流行的目标检测算法相比取得了最好的性能，与原始YOLOv8模型相比，改进后的模型mAP达到89.02%,mAP50∶90增加了4.08%,能够高效准确地识别矿带分界点，展现了该模型在矿带分界点检测领域中的巨大潜力.
Abstract：The traditional shaking table beneficiation methods relies on manual adjustment of the ore-receiving plate position by workers on their experience, which leads to higher labor costs and unstable results.The development of deep learning technology provides the possibility for the intelligent application of shaking tables.Since its release, the YOLOv8 object detection model has been widely recognized in the field of object detection for its high detection accuracy and robustness.However, YOLOv8 has limitations when dealing with complex scenes, such as insufficient shallow feature extraction and insufficient channel expression ability of higher-order feature maps.To address these problems, this paper proposes the Low Level Attention(LLA) and the High Level Attention(HLA) modules, which enhance the expression ability of low and high order feature maps, respectively, and reduce the number of model parameters by introducing the C2f-faster module to achieve a lightweight model.Experiments show that compared to five popular object detection algorithms, our model achieves the best performance.Compared to the original YOLOv8 model, the improved model's mAP reaches 89.02%,and the mAP50∶90 increases by 4.08%,efficiently and accurately identifying the boundary point of the ore belt.This demonstrates the great potential of this model in the field of demarcation point detection.
[1] 程建忠，车丽萍.中国稀土资源开采现状及发展趋势[J].稀土，2010,31(2):65-69+85.CHENG J Z,CHE L P.Current mining situation and potential development of rare earth in China[J].Chinese Rare Earths,2010,31(2):65-69+85.
[2] IZERDEM D,ERGUN S L.Investigation of the effects of particle size on the performance of classical gravity concentration equipment[J].Mineral Processing and Extractive Metallurgy Review,2024,45(3):155-172.
[3] 谭明，沈政昌，杨义红.矿物分选装备技术研究进展[J].绿色矿山，2024(1):85-93.TAN M,SHEN Z C,YANG Y H.Research progress of mineral processing equipment technology[J].Journal of Green Mine,2024(1):85-93.
[4] 丁建军，白飞燕，任学禹，等.基于浮选泡沫图像识别的精煤灰分预测系统[J].选煤技术，2022,50(4):89-93.DING J J,BAI F Y,REN X Y,et al.The flotation froth image recognition-based concentrate ash prediction system[J].Coal Preparation Technology,2022,50(4):89-93.
[5] 刘惠中，芮作为，朱合钧，等.基于改进YOLOv5算法的选矿摇床矿带分离点目标检测识别研究[J].有色金属科学与工程，2025,16(1):115-124.LIU H Z,RUI Z W,ZHU H J,et al.Recognition on ore zone separation points target detection and identification in mineral processing shaking table based on the improved YOLOv5 algorithm[J].Nonferrous Metals Science and Engineering,2025,16(1):115-124.
[6] 和丽芳，黄斌，黄宋魏，等.一种基于改进海鸥算法的锡矿摇床矿带分带图像分割法：CN116645331A[P].2023-08-25.
[7] 杨林顺，刘航涛.基于深度残差网络的煤泥浮选泡沫图像分类方法研究[J].煤炭技术，2023,42(7):226-229.YANG L S,LIU H T.Study on forth image classification of coal flotation based on deep residual network[J].Coal Technology,2023,42(7):226-229.
[8] 谢涛，余子昭，赵程，等.基于Segnet语义分割网络初至拾取方法及应用[J/OL].地球物理学进展，2024:1-12[2024-07-05].https://kns.cnki.net/kcms/detail/11.2982.P.20240704.1358.002.html.XIE T,YU Z Z,ZHAO C,et al.Segnet-based semantic segmentation network first-to-pickup method and application[J/OL].Progress in Geophysics,2024:1-12[2024-07-05].https://kns.cnki.net/kcms/detail/11.2982.P.20240704.1358.002.html.
[9] 张梦妮，王祎鸣，吴勇剑.基于改进Mask R-CNN的船载地波雷达目标检测方法[J/OL].海洋科学进展，1-13[2025-01-25].http://kns.cnki.net/kcms/detail/37.1387.P.20240703.1009.002.html.ZHANG M N,WANG Y M,WU Y J.Target detecyion method of shipborne surface wave radar based on improved Mask R-CNN[J/OL].Advances in Marine Science,1-13[2025-01-25].http://kns.cnki.net/kcms/detail/37.1387.P.20240703.1009.002.html.
[10] YOU K S,WEN C Y,LIU H Z.Research on intelligent implementation of the beneficiation process of shaking table[J].Minerals Engineering,2023,199:108108.
[11] 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),June 27-30,2016,Las Vegas,NV,USA.IEEE,2016:779-788.
[12] 米增，连哲.面向通用目标检测的YOLO方法研究综述[J].计算机工程与应用，2024,60(21):38-54.MI Z,LIAN Z.Review of YOLO methods for universal object detection[J].Computer Engineering and Applications,2024,60(21):38-54.
[13] VARGHESE R,SAMBATH M.YOLOv8:A novel object detection algorithm with enhanced performance and robustness[C]//2024 International Conference on Advances in Data Engineering and Intelligent Computing Systems (ADICS),April,18-19,2024.Chennai,India.IEEE,2024:1-6.
[14] CHEN J R,KAO S H,HE H,et al.Run,don’t walk:Chasing higher FLOPS for faster neural networks[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR),Vancouver,BC,Canada.IEEE,2023:12021-12031.
[15] LUO W J,LI Y J,URTASUN R,et al.Understanding the effective receptive field in deep convolutional neural networks[C]//Proceedings of the 30th International Conference on Neural Information Processing Systems,December 5-10,2016,Barcelona,Spain.ACM,2016:4905-4913.
[16] LIN T Y,DOLLAR P,GIRSHICK R,et al.Feature pyramid networks for object detection[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),July 21-26,2017,Honolulu,HI.IEEE,2017:936-944.
[17] LAWAL O M.YOLOv5-LiNet:A lightweight network for fruits instance segmentation[J].PLoS One,2023,18(3):e0282297.
[18] 邝先验，程福军，吴翠琴，等.基于改进YOLOv7-tiny的高效轻量遥感图像目标检测方法[J].电子测量与仪器学报，2024,38(7):22-33.KUANG X Y,CHENG F J,WU C Q,et al.Efficient and lightweight target detection method for remote sensing images based on improved YOLOv7-tiny[J].Journal of Electronic Measurement and Instrumentation,2024,38(7):22-33.
[19] 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(6):2207-2216.XU Z Q,Lü Z Q,WANG W D,et al.Machine vision recognition method and optimization for intelligent separation of coal and gangue[J].Journal of China Coal Society,2020,45(6):2207-2216.
[20] 赵然磊，杨留栓，徐晓，等.基于XGBoost算法的火山岩岩性识别方法与研究[J].地球物理学进展，2025,40(2):646-657.ZHAO R L,YANG L S,XU X,et al.Lithology identification method and research of volcanic rock based on XGBoost algorithm[J].Progress in Geophysics,2025,40(2):646-657.
[21] 张瀚文，曹维娟，罗刚银，等.基于改进型U-Net的变色油墨血浆判别模型[J].南京医科大学学报(自然科学版),2024,44(9):1179-1189.ZHANG H W,CAO W J,LUO G Y,et al.A plasma discrimination model for color changing ink based on improved U-Net[J].Journal of Nanjing Medical University(Natural Sciences),2024,44(9):1179-1189.
基本信息:
DOI：10.16112/j.cnki.53-1223/n.2026.02.522
中图分类号:TD455.2;TP183;TP391.41
引用信息:
[1]邵平,刘丹,余龙舟,等.基于改进YOLOv8的摇床矿带分界点目标检测算法[J].昆明理工大学学报(自然科学版),2025,50(05):45-55.DOI:10.16112/j.cnki.53-1223/n.2026.02.522.
基金信息:
国家自然科学基金项目（52464028）