Object Detection of Steel Mesh Binding Point Using FNB-YOLOv5

doi:10.16183/j.cnki.jsjtu.2024.121

Abstract

Abstract:

To address the problems of low accuracy and slow detection speed in existing target detection algorithms used by rebar-binding robots for identifying binding points, an improved rebar mesh binding-point detection method named FNB-YOLOv5, was proposed based on enhancements to YOLOv5. First, a dataset of rebar grid intersections was created through image acquisition. Then, the lightweight FasterNet backbone was incorporated into the YOLOv5 network to enhance feature extraction while reducing network complexity. Next, the BiFormer attention mechanism was introduced into key network components to improve the accuracy of feature extraction. Considering that small-sized targets dominate the detection task, the NWD loss function was used for normalization to optimize the localization accuracy of binding-point detection. Finally, an improved F-global feature pyramid network (F-GFPN) feature fusion module was devised to enhance feature interaction and improve computational performance by incorporating skip connections and cross-scale connections. Comparative experiments with different models show that the proposed method achieves a precision of 99.82%, a recall of 99.10%, and a mean average precision of 98.64%, which represent an increase of 2.9 percentage points, 1.53 percentage points, and 1.63 percentage points over the original model, respectively. The frame per second (FPS) reaches 44.8, an increase of 4.1 compared with the original model, while the size of weight model is reduced by 2.93 MB. Experimental results demonstrate that the improved FNB-YOLOv5 model achieves higher accuracy and real-time performance on the rebar binding point dataset, providing technical support for the development of rebar binding robots in the construction industry.

Key words: binding point of steel mesh, YOLOv5, small target detection, image processing

CLC Number:

TP391.4

LI Zixuan, ZHAO Zhigang, ZHANG Zeyu, JIE Junjie, CHENG Ruiqiang. Object Detection of Steel Mesh Binding Point Using FNB-YOLOv5[J]. Journal of Shanghai Jiao Tong University, 2026, 60(5): 762-775.

Figures/Tables 17

Fig.1

Fig.2

Fig.3

Fig.4

Fig.5

Fig.6

Fig.7

Fig.8

Tab.1

Fig.9

Tab.2

Tab.3

Fig.10

Fig.11

Fig.12

Fig.13

Tab.4

References 31

[1]	CARDNO C A. Robotic rebar-tying system uses artificial intelligence[J]. Civil Engineering, 2018, 88(1): 38-39.
[2]	Future Robotics Technology Center. T-iROBO rebar[EB/OL]. (2022-08-01) [2024-03-01]. https://www.furo.org/en/works/t_irobo_rebar/t_irobo_rebar.html.
[3]	KEN R. About TOMOROBO[EB/OL]. (2022-08-01) [2024-03-01]. https://en.kenrobo-tech.com/tom-orobo/.
[4]	FENG R, JIA Y, WANG T, et al. Research on the system design and target recognition method of the rebar-tying robot[J]. Buildings, 2024, 14(3): 838. doi: 10.3390/buildings14030838 URL
[5]	JIN J, ZHANG W, LI F, et al. Robotic binding of rebar based on active perception and planning[J]. Automation in Construction, 2021, 132: 103939. doi: 10.1016/j.autcon.2021.103939 URL
[6]	董国梁, 张雷, 辛山. 基于深度学习的钢筋绑扎机器人目标识别定位[J]. 电子测量技术, 2022, 45(11): 35-44.
	DONG Guoliang, ZHANG Lei, XIN Shan. Target identification and location of steel binding robot based on deep learning[J]. Electronic Measurement Technology, 2022, 45(11): 35-44.
[7]	WOO S, PARK J, LEE J Y, et al. CBAM: Convolutional block attention module[C]// 2018 15th European Conference on Computer Vision (ECCV). Munich, Germany: ECCV, 2018: 3-19.
[8]	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. Salt Lake City, UT, USA: CVPR, 2018: 4510-4520.
[9]	成彬, 邓磊. 融合深度学习与特征投影曲线的钢筋绑扎节点检测方法[J]. 计算机辅助设计与图形学学报, 2024, 36(8): 1272-1285.
	CHENG Bin, DENG Lei. Rebar binding crosspoints detection method based on deep learning and feature projection curve[J]. Journal of Computer-Aided Design & Computer Graphics, 2024, 36(8): 1272-1285.
[10]	肖恭轩. 视觉协作下机械臂PC构件钢筋绑扎技术研究[D]. 福建: 福建工程学院, 2023.
	XIAO Gongxuan. Research on robotic arm to PC components steel binding technology with visual collaboration abstract[D]. Fujian: Fujian University of Technology, 2023.
[11]	ZHENG Z, WANG P, REN D, et al. Enhancing geometric factors in model learning and inference for object detection and instance segmentation[J]. IEEE Transactions on Cybernetics, 2022, 52(8): 8574-8586. doi: 10.1109/TCYB.2021.3095305 URL
[12]	GIRSHICK R. Fast R-CNN[C]// 2015 IEEE International Conference on Computer Vision. Santiago, Chile: ICCV, 2015: 1440-1448.
[13]	HE K, GKIOXARI G, DOLLAR P, et al. Mask R-CNN[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(2): 386-397. doi: 10.1109/TPAMI.2018.2844175 pmid: 29994331
[14]	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. doi: 10.1109/TPAMI.2016.2577031 pmid: 27295650
[15]	REDMON J, DIVVALA S K, GIRSHICK R B, et al. You only look once: Unified, real-time object detection[C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, NV, USA: CVPR, 2016: 779-788.
[16]	REDMON J, FARHADI A. YOLO9000: Better, faster, stronger[C]// 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, HI, USA: CVPR, 2017: 6517-6525.
[17]	REDMON J, FARHADI A. YOLOv3:An incremental improvement[EB/OL]. (2018-04-09)[2024-03-01]. https://arxiv.org/abs/1804.02767.
[18]	GE Z, LIU S, WANG F, et al. YOLOX:Exceeding YOLO series in 2021[EB/OL]. (2021-07-18)[2024-03-01]. https://arxiv.org/abs/2107.08430.
[19]	QIAO S, CHEN L C, YUILLE A. DetectoRS: Detecting objects with recursive feature pyramid and switchable atrous convolution[C]// 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Nashville, TN, USA: CVPR, 2021: 10208-10219.
[20]	LI C, LI L, JIANG H, et al. YOLOv6:A single-stage object detection framework for industrial applications[EB/OL]. (2022-09-07)[2024-03-01]. https://arxiv.org/abs/2209.02976.
[21]	WANG C Y, BOCHKOVSKIY A, LIAO H, et al. YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C]// 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Vancouver, BC, Canada: CVPR, 2023: 7464-7475.
[22]	LIU W, ANGUELOV D, ERHAN D, et al. SSD: Single shot MultiBox detector[C]// 14th European Conference on Computer Vision (ECCV). Amsterdam, Netherlands: ECCV, 2016: 21-37.
[23]	CHEN J, 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: CVPR, 2023: 12021-12031.
[24]	ZHU L, WANG X, KE Z, et al. BiFormer: Vision transformer with bi-level routing attention[C]// 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Vancouver, Canada: CVPR, 2023: 10323-10333.
[25]	WANG J, XU C, YANG W, et al. A normalized Gaussian Wasserstein distance for tiny object detection[EB/OL]. (2022-07-14)[2024-03-01]. https://arxiv.org/abs/2110.13389.
[26]	XU X, JIANG Y, CHEN W, et al. DAMO-YOLO: A report on real-time object detection design[EB/OL]. (2022-11-23)[2024-03-01]. https://arxiv.org/abs/2211.1544.
[27]	薛昂, 姜恩宇, 张文涛, 等. 基于窗口自注意力网络与YOLOv5融合的输电线路通道异物检测[J]. 上海交通大学学报, 2025, 59(3): 413-423. doi: 10.16183/j.cnki.jsjtu.2023.301
	XUE Ang, JIANG Enyu, ZHANG Wentao, et al. Detection of foreign bodies in transmission line channels based on the fusion of Swin Transformer and YOLOv5[J]. Journal of Shanghai Jiao Tong University, 2025, 59(3): 413-423.
[28]	LIU S, QI L, QIN H, et al. Path aggregation network for instance segmentation[C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA: CVPR, 2018: 8759-8768.
[29]	HOWARD A G, ZHU M, CHEN B, et al. MobileNets: Efficient convolutional neural networks for mobile vision applications[EB/OL]. (2017-04-17)[2024-03-01]. https://arxiv.org/abs/1704.04861.
[30]	MA N, ZHANG X, ZHENG H T, et al. ShuffleNet V2: Practical guidelines for efficient CNN architecture design[C]// 15th European Conference on Computer Vision (ECCV). Munich, Germany: ECCV, 2018: 122-138.
[31]	HAN K, WANG Y, TIAN Q, et al. GhostNet: More features from cheap operations[C]// 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle, WA, USA: CVPR, 2020: 1577-1586.

参数	取值
epoch	60
优化器	Adam
lr0	0.01
(β₁, β₂)	(0.9, 0.999)
Batch_size	12

基础模型	FasterNet	BiFormer	NWD	F-GFPN	P/%	mAP@0.5/%	mAP@0.5∶0.95/%	FPS/(帧·s^-1)
√					96.92	97.01	63.05	40.7
√	√				96.55	97.84	63.11	37.9
√		√			97.31	96.57	62.77	40.3
√			√		97.12	97.57	64.29	43.2
√				√	97.84	97.97	66.90	38.5
√	√	√	√	√	99.82	98.64	66.79	44.8

对比模型	P/%	R/%	mAP@0.5/%	Weight/MB	GFLOPs	FPS/(帧·s^-1)
Faster R-CNN^[4]	66.91	—	—	—	—	18.45
VGG16-SSD^[7]	86.30	88.00	—	85.70	—	—
YOLOv5s	96.92	97.57	97.01	16.78	1.66	40.7
YOLOv5s-VOC	96.65	97.66	96.83	13.84	1.58	45.5
YOLOv5s6	96.68	97.59	96.82	14.78	1.62	43.3
YOLOv5l	98.88	99.31	98.92	89.51	11.39	11.1
YOLOv5m	98.46	99.08	98.41	40.63	5.02	22.3
YOLOv8n	98.07	97.82	98.29	17.25	1.87	35.6
FNB-YOLOv5	99.82	99.10	98.64	13.85	1.63	44.8

绑扎点识别模型	已绑扎点识别数	未绑扎点识别数	已绑扎点识别率/%	未绑扎点识别率/%	整体识别率/%
YOLOv5s	4	5	100.0	62.5	75.0
YOLOv5s-VOC	4	6	100.0	75.0	83.3
YOLOv5s6	4	7	100.0	87.5	91.7
YOLOv5m	4	8	100.0	100.0	100.0
YOLOv8n	4	7	100.0	87.5	91.7
FNB-YOLOv5 (本文)	4	8	100.0	100.0	100.0