融合超分辨率重建的YOLOv7煤矸石识别模型

doi:10.3778/j.issn.1002-8331.2404-0186

摘要/Abstract

摘要： 在洗煤场景下拍摄的影像中煤矸石目标光照低且分布不均匀，同时在洗煤过程中目标存在遮挡、过曝或逆光现象，使得边界和纹理较模糊，提取特征过程中纹理特征容易丢失，造成误检。因此，对不均匀低光照的煤矸石目标检测方法进行研究，提出一种融合超分辨率重建的煤矸石识别模型。对于图像中的小目标煤矸石，利用超分辨率重建技术将分辨率提高后再进一步利用YOLOv7模型检测目标以提升整体检测效果。在超分辨重建模块中加入煤矸石锐化功能模块，提升煤矸石目标边缘清晰度，以保证在特征提取时通过浅层特征层能更好地提取煤矸石目标轮廓、形状等信息。在构建的煤矸石数据集上进行煤矸石目标检测，实验结果表明，改进后算法在煤、矸石检测任务中平均精度分别达到99.52%与98.84%。且模型在非均匀光照、粉尘干扰场景保持识别稳定性，煤、矸石漏检率分别降低18和17个百分点。改善了小目标与纹理模糊目标识别困难问题，为煤矸石识别提供技术参考。

关键词: 煤矸石检测, 超分辨率重建, 小目标检测, 图像识别

Abstract: In the image taken in the coal washing scene, the target illumination of coal gangue is low and the distribution is uneven. Meanwhile, the target is blocked, overexposed or backlit during the coal washing process, which makes the boundary and texture fuzzy, and the texture features are easily lost during the feature extraction process, resulting in false detection. Therefore, the detection method of coal gangue with non-uniform low illumination is studied, and a coal gangue recognition model with fusion super-resolution reconstruction is proposed. For the small target coal gangue in the image, super resolution reconstruction technology is used to improve the resolution, and then YOLOv7 model is further used to detect the target to improve the overall detection effect. The coal gangue sharpening function module is added into the super-resolution reconstruction module to improve the clarity of the coal gangue target edge, so as to ensure that the outline and shape of the coal gangue target can be better extracted through the shallow feature layer during feature extraction. The coal gangue target detection is carried out on the constructed coal gangue dataset. The proposed algorithm achieves mean average precision of 99.52% and 98.84% for coal and gangue detection. Moreover, it maintains robust recognition under non-uniform illumination and dust interference, reducing the miss-detection rates of coal and gangue by 18 and 17 percentage points, respectively, which improves the difficulty of identifying small targets and fuzzy texture targets, and provides technical reference for coal gangue identification.

Key words: coal gangue detection, super resolution reconstruction, small target detection, image recognition

李娜, 秦昆德, 李想, 张澳迪. 融合超分辨率重建的YOLOv7煤矸石识别模型[J]. 计算机工程与应用, 2025, 61(15): 343-352.

LI Na, QIN Kunde, LI Xiang, ZHANG Aodi. YOLOv7 Coal Gangue Recognition Model Fused with Super-Resolution Reconstruction[J]. Computer Engineering and Applications, 2025, 61(15): 343-352.

参考文献

[1] 马宏伟. 煤矿机电装备智能化[J]. 西安科技大学学报, 2020, 40(5): 748.
MA H W. Intelligent coal mine electromechanical equipment[J]. Journal of Xi’an University of Science and Technology, 2020, 40(5): 748.
[2] 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.
[3] BESSINGER S L, NELSON, M G Remnant roof coal thickness measurement with passive gamma ray instruments in coal mines[J]. IEEE Transactions on Industry Applications, 1993, 29(3): 562-565.
[4] 马宪民, 蒋勇, 卜祥莉. 基于图像处理的煤矸石自动分选系统的研究[C]//2003年中国智能自动化会议论文集(下册), 2003: 403-406.
MA X M, JIANG Y, PU X L. Study of automatic selection system of waste rock from coal bulk based on image pro-cessing[C]//Proceedings of the 2003 China Intelligent Automation Conference(Volume 2), 2003: 403-406.
[5] 薛祯也. 基于深度学习的煤矸石识别方法研究[D]. 西安: 西安科技大学, 2019.
XUE Z Y. Research on coal gangue identification based on deep learning[D]. Xi’an: Xi’an University of Science and Technology, 2019.
[6] 王莉, 于国防, 沈慧宇, 等. 基于CNN卷积神经网络的煤矸石自动分选研究[J]. 江苏建筑职业技术学院学报, 2019, 19(4): 35-39.
WANG L, YU G F, SHEN H Y, et al. Study onautomatic separation of coal gangue based on CNN convolutional neural network[J]. Jiangsu Vocational Institute of Architectural Technology, 2019, 19(4): 35-39.
[7] LV Z Q, WANG W D, XU Z Q, et al. Cascade network for detection of coal and gangue in the production context[J]. Powder Technology, 2021, 377: 361-371.
[8] DING Z H, CHEN G D, WANG Z, et al. A real-time multilevel fusion recognition system for coal and gangue based on near-infrared sensing[J]. IEEE Access, 2020, 8: 178722-178732.
[9] 吕旻姝. 基于深度学习的煤矸石图像识别研究[D]. 淮南: 安徽理工大学, 2021.
LV M S. Research on coal gangue image recognition based on deep learning[D]. Huainan: Anhui University of Science and Technology, 2021.
[10] 苏玲玲. 基于深度学习的煤矸石识别方法研究[D]. 西安: 西安科技大学, 2018.
SU L L. Research on coal gangue identification based on deep learning[D]. Xi’an: Xi’an University of Science and Technology, 2018.
[11] 赵学军, 李建. 一种基于深度学习的煤矸石检测方法[J]. 矿业科学学报, 2021, 6(6): 730-736.
ZHAO X J, LI J. A method of coal gangue detection based on deep learning[J]. Journal of Mining Science and Technology, 2021, 6(6): 730-736.
[12] LI M, HE X L, YUAN Y X, et al. Multiple factors influence coal and gangue image recognition method and experimental research based on deep learning[J]. International Journal of Coal Preparation and Utilization, 2023, 43(8): 1411-1427.
[13] XU S Y, ZHOU Y J, HUANG Y R, et al. YOLOv4-tiny-based coal gangue image recognition and FPGA implementation[J]. Micromachines, 2022, 13(11): 1983.
[14] 汝洪芳, 张冬冬. YOLOv5检测煤矸石的改进方法[J]. 黑龙江科技大学学报, 2021, 31(6): 818-823.
RU H F, ZHANG D D. Coal gangue detection method based on improved YOLOv5[J]. Journal of Heilongjiang University of Science and Technology, 2021, 31(6): 818-823.
[15] DONG C, LOY C C, HE K M, et al. Learning a deep convolutional network for image super-resolution[C]//Proceedings of the European Conference on Computer Vision. Cham: Spr-inger International Publishing, 2014: 184-199.
[16] SHI W Z, CABALLERO J, HUSZáR F, et al. Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2016: 1874-1883.
[17] KIM J, LEE J K, LEE K M. Accurate image super-resolution using very deep convolutional networks[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2016: 1646-1654.
[18] KIM S, KANG I, KWAK N. Semantic sentence matching with densely-connected recurrent and co-attentive information[C]//Proceedings of the AAAI Conference on Artificial Intelligence, 2019: 6586-6593.
[19] HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2016: 770-778.
[20] ZHANG Y L, TIAN Y P, KONG Y, et al. Residual dense network for image super-resolution[C]//Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2018: 2472-2481.
[21] LEDIG C, THEIS L, HUSZáR F, et al. Photo-realistic single image super-resolution using a generative adversarial network[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 105-114.
[22] ZHANG K, LIANG J Y, VAN GOOL L, Designing a practical degradation model for deep blind image super-resolution[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2021: 4791-4800.
[23] WANG X T, YU K, WU S X, et al. ESRGAN: enhanced super-resolution generative adversarial networks[C]//Proceedings of the European Conference on Computer Vision. Cham: Springer International Publishing, 2019: 63-79.
[24] JI H, GAO Z, MEI T C, et al. Vehicle detection in remote sensing images leveraging on simultaneous super-resolution[J]. IEEE Geoscience and Remote Sensing Letters, 2019, 17(4): 676-680.
[25] RABBI J, RAY N, SCHUBERT M, et al. Small-object dete-ction in remote sensing images with end-to-end edge-enhanced GAN and object detector network[J]. Remote Sensing, 2020, 12(9): 1432.
[26] HE S T, ZOU H X, WANG Y Q, et al. ShipSRDet: an end-to-end remote sensing ship detector using super-resolved feature representation[C]//Proceedings of the 2021 IEEE International Geoscience and Remote Sensing Symposium. Piscataway: IEEE, 2021: 3541-3544.
[27] 张天霖, 逄征, 陈红珍, 等. 面向舰船目标识别的遥感图像超分辨率重建[J]. 计算机工程与应用, 2024, 60(13): 190-199.
ZHANG T L, PANG Z, CHEN H Z, et al. Remote sensing image super-resolution reconstruction method for ship target recognition[J]. Computer Engineering and Applications, 2024, 60(13): 190-199.
[28] 王文瑾, 游子绎, 邵历江, 等. 融合超分辨率重建的YOLOv5松枯死木识别模型[J]. 农业工程学报, 2023, 39(5): 137-145.
WANG W J, YOU Z Y, SHAO L J, et al. Recognition of dead pine trees using YOLOv5 by super-resolution reconstruction[J]. Transactions of the Chinese Society of Agricultural Engineering, 2023, 39(5): 137-145.
[29] WANG C Y, BOCHKOVSKIY A, LIAO H M. YOLOv7: trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C]//Proceedings of the 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2023: 7464-7475.