|本期目录/Table of Contents|

[1]黄 灿,叶安源,易浩洋,等. 基于改进YOLOv8算法的核环境水下机器人焊缝识别 [J].武汉工程大学学报,2025,47(03):325-330,354.[doi:10.19843/j.cnki.CN42-1779/TQ.202504018]
　HUANG Can,YE Anyuan,YI Haoyang,et al. Weld seam recognition in nuclear environments by underwater robot based on improved YOLOv8 algorithm [J].Journal of Wuhan Institute of Technology,2025,47(03):325-330,354.[doi:10.19843/j.cnki.CN42-1779/TQ.202504018]
点击复制

基于改进YOLOv8算法的核环境水下机器人焊缝识别
(/HTML)
分享到：

《武汉工程大学学报》[ISSN:1674-2869/CN:42-1779/TQ]

卷:
47
期数:
2025年03期
页码:
325-330,354
栏目:
智能制造
出版日期:
2025-06-30

文章信息/Info

Title:
Weld seam recognition in nuclear environments by underwater robot based on improved YOLOv8 algorithm
文章编号:
1674 - 2869（2025）03 - 0325 - 06
作者:
1. 中核运维技术有限公司，浙江 杭州 310000；
2. 武汉工程大学机电工程学院，湖北 武汉 430205
Author(s):
1. CNNC Operation and Maintenance Technology Co., Ltd, Hangzhou 310000, China;
2. School of Mechanical and Electrical Engineering, Wuhan Institute of Technology, Wuhan 430205, China
关键词:
Keywords:
分类号:
TP241
DOI:
10.19843/j.cnki.CN42-1779/TQ.202504018
文献标志码:
A
摘要:
针对核环境水下机器人焊缝识别困难的情况，提出了一种基于改进YOLOv8算法的焊缝识别模型。首先，通过引入水下随机气泡噪声，提高了焊缝识别模型的泛化能力。其次，通过引入多维协作注意力机制和双向特征金字塔网络，在保证额外参数量和计算量相近的基础上，提高了水下目标的模型检测精度。为验证改进模型性能，在水下环境中进行了焊缝识别实验及算法消融实验，结果表明，基于改进YOLOv8算法的焊缝识别模型相比于改进前，平均预测精度上升13.9%，且未引入额外计算量，有效验证了改进模型的鲁棒性和有效性。核作业水下机器人通过算法自动识别焊缝并完成路径跟踪，显著减少了人工检测的风险和操作难度，有效避免了工作人员暴露于辐射环境，同时大幅缩短了核电站大修周期，降低了潜在的经济损失。
Abstract:
To address the difficulty of weld seam recognition by underwater robots in nuclear environment, this study proposed an improved weld seam recognition model based on YOLOv8 algorithm. First, the generalization ability of the weld seam recognition model was improved by introducing underwater random bubble noise. Then, by introducing multidimensional collaborative attention and bidirectional feature pyramid network, the model’s detection accuracy for underwater targets was effectively improved while maintaining similar amount of extra parameters and calculation. Finally, weld seam recognition experiments and algorithm ablation studies were conducted in an underwater environment. The results demonstrated that compared with the original algorithm, the weld recognition model based on improved YOLOv8 achieved a 13.9% increase in average precision without introducing additional computational cost, effectively validating the improved model’s robustness and effectiveness. By enabling the underwater robot to automatically identify weld seams and perform path tracking, the algorithm proposed significantly reduces the risks and operational difficulties associated with manual inspection, effectively avoids personnel exposure to radiation, substantially shortens the overhaul period of nuclear power plants, and reduces potential economic losses.

参考文献/References:

[1] 李浩. 基于视觉的核电站水下焊接机器人焊缝识别与路径跟踪[D]. 哈尔滨：哈尔滨工业大学, 2018.
[2] 苏志威, 黄子涵, 邱发生, 等. 基于改进YOLOv8的航空铝合金焊缝缺陷检测方法[J]. 航空动力学报, 2024, 39(6): 20230414.
[3] 张润梅, 潘晨飞, 陈梓华, 等. 改进YOLOv8的焊缝表面缺陷检测算法[J]. 光电工程,2025,52(3): 240296.
[4] 於双飞, 卓伟, 王保华, 等. 基于改进核相关滤波的焊缝特征点鲁棒识别算法[J]. 激光与光电子学进展, 2025，62(14): 4.
[5] 王仕仙, 陈绪兵. 焊接轨迹跟踪控制中的深度视觉研究进展[J]. 武汉工程大学学报,2023,45(4):378-383，417.
[6] SAZONOVA S A, NIKOLENKO S D, OSIPOV A A, et al. Weld defects and automation of methods for their detection[J]. Journal of Physics: Conference Series. 2021, 1889(2): 022078.
[7] ZUO F Y, LIU J H, FU M R, et al. An end-to-end intelligent welding defect detection system for low-quality X-ray images with adaptive progressive learning[J]. Expert Systems With Applications, 2025, 279:127428.
[8] NADDAF-SH A M, BABURAO V S, ZARGARZA-DEH H. Leveraging segment anything model (SAM) for weld defect detection in industrial ultrasonic B-scan images[J]. Sensors, 2025, 25(1): 277.
[9] 陈圣峰. 基于结构光视觉传感的机器人GMAW智能焊缝识别与定位方法研究[D]. 长沙：湖南大学,2022.
[10] 黄瑜豪, 曾祥进, 冯崧. 面向边缘设备的轻量级OpenPose姿态检测模型研究[J]. 武汉工程大学学报,2024,46(4):424-430.
[11] 龚律凯, 彭伊丽, 陈绪兵, 等. 基于改进U-Net算法的焊缝特征识别研究[J]. 现代制造工程, 2024(11): 18-25.
[12] 严蓬辉, 陈绪兵, 彭伊丽, 等. 基于改进YOLOv5s的激光软钎焊焊点缺陷检测算法[J]. 激光与光电子学进展, 2024, 61(8): 0812005.
[13] 乐恒韬, 赵康康, 吴松林, 等. 基于LSTM网络的机器人异空间手眼标定方法[J]. 武汉工程大学学报, 2024, 46(5): 574-578.
[14] RUIZ-PONCE P, ORTIZ-PEREZ D, GARCIA-RODRIGUEZ J, et al. POSEIDON: a data augmentation tool for small object detection datasets in maritime environments[J]. Sensors, 2023, 23(7): 3691.
[15] YU Y, ZHANG Y, CHENG Z Y, et al. MCA: multidimensional collaborative attention in deep convolutional neural networks for image recognition[J]. Engineering Applications of Artificial Intelli-gence, 2023, 126: 107079.
[16] TAN M X, PANG R M, LE Q V. EfficientDet: scalable and efficient object detection[C]// 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ: IEEE, 2020: 10781-10790.

相似文献/References:

备注/Memo

备注/Memo:
收稿日期：2025-04-24
基金项目：国家自然科学基金 (52205536)
作者简介：黄 灿，高级工程师。Email：huangcan@cnnp.com.cn
*通信作者：彭伊丽，博士，讲师。Email：21040301@wit.edu.cn
引文格式：黄灿，叶安源，易浩洋，等. 基于改进YOLOv8算法的核环境水下机器人焊缝识别［J］. 武汉工程大学学报，2025，47（3）：325-330，354.
更新日期/Last Update: 2025-07-09