基于浅层特征抑制的油气管道小缺陷检测网络

doi:10.12422/j.issn.1672-6952.2024.06.011

辽宁石油化工大学学报 ›› 2024, Vol. 44 ›› Issue (6): 81-88.DOI: 10.12422/j.issn.1672-6952.2024.06.011

基于浅层特征抑制的油气管道小缺陷检测网络

郝鹏程¹(), 郎宪明¹(), 郭晓庆²

^1.辽宁石油化工大学信息与控制工程学院，辽宁抚顺 113001
^2.中航沈飞民用飞机有限责任公司，辽宁沈阳 110000

收稿日期:2024-08-07 修回日期:2024-09-04 出版日期:2024-12-25 发布日期:2024-12-24
通讯作者: 郎宪明
作者简介:郝鹏程（1999⁃），男，硕士研究生，从事管道缺陷识别检测和漏磁图像处理方面的研究；E⁃mail：haopengcheng@stu.lnpu.edu.cn。
基金资助:
国家自然科学基金项目(62073158);辽宁省自然科学基金计划面上项目(2023?MS?289);辽宁省教育厅基本科研项目(JYTMS20231441);辽宁省应用基础研究计划项目(2023JH26/10300013);辽宁振兴人才基金项目(XLYC220316)

Detection Network for Small Defects in Oil and Gas Pipelines Based on Shallow Feature Suppression

Pengcheng HAO¹(), Xianming LANG¹(), Xiaoqing GUO²

^1.School of Information and Control Engineering，Liaoning Petrochemical University，Fushun Liaoning 113001，China
^2.AVIC SAC Commercial Aircraft Company Limited，Shenyang Liaoning 110000，China

Received:2024-08-07 Revised:2024-09-04 Published:2024-12-25 Online:2024-12-24
Contact: Xianming LANG

摘要/Abstract

摘要：

针对在漏磁检测中石油和天然气管道中的小缺陷样本稀缺及检测精度不佳的问题，提出了基于浅层特征抑制的油气管道小缺陷检测网络。首先，利用对抗生成网络并融入先验知识，以生成高质量小缺陷样本。然后，在特征提取过程中引入缺陷特征抑制模块，在浅层金字塔特征中抑制大缺陷语义从而增强小缺陷特征。最后，多尺度注意力变换器（Transformer）充分利用特征图像的局部细节和全局信息提高管道缺陷检测准确率。实验结果表明，该模型检测的准确率为95.1%，比现有的Faster R?CNN等方法的平均值高7.8%。

关键词: 小缺陷检测, 漏磁图像, 特征金字塔, 卷积神经网络, 注意力机制

Abstract:

To address the issues of scarce small defect samples and poor detection accuracy in magnetic flux leakage (MFL) testing of oil and gas pipelines, this paper proposes a small defect detection network for oil and gas pipelines on the basis of shallow feature suppression. First, an adversarial generative network is utilized, which incorporates prior knowledge to generate high?quality small defect samples. Subsequently, a defect feature suppression module is introduced during the feature extraction process, which suppresses the semantics of large defects in shallow pyramid features, thereby enhancing the features of small defects. Finally, a multi?scale Transformer is employed to fully leverage the local details and global information of the feature images, improving the accuracy of pipeline defect detection. The experimental results demonstrate that the accuracy of this model is 95.1%, which is 7.8% higher than the average value of existing faster R?CNN and other methods methods.

Key words: Small defect detection, Magnetic flux leakage, Pyramid feature fusion, CNN, Attention mechanisms

中图分类号:

TE88

郝鹏程, 郎宪明, 郭晓庆. 基于浅层特征抑制的油气管道小缺陷检测网络[J]. 辽宁石油化工大学学报, 2024, 44(6): 81-88.

Pengcheng HAO, Xianming LANG, Xiaoqing GUO. Detection Network for Small Defects in Oil and Gas Pipelines Based on Shallow Feature Suppression[J]. Journal of Liaoning Petrochemical University, 2024, 44(6): 81-88.

图/表 8

图1 管道漏磁检测器

Fig.1 Magnetic flux leakage detector for pipelines

图2 M?GAN网络结构

Fig.2 M?GAN network architecture

图3 特征抑制语义提取

Fig.3 Feature suppression semantic extraction

图4 管道漏磁小缺陷检测实验环境及漏磁图像

Fig.4 MFL pipeline small defect detection experiment and MFL image

图5 漏磁模型与生成图像（a）漏磁模型（b）生成图像

Fig.5 MFL model and generated image

图6 漏磁图像缺陷检测结果及置信度

Fig.6 MFL image defect detection results and confidence level

表1 管道漏磁缺陷检测结果对比

Table 1 Comparison of pipeline mfl defect detection results

方法	P/%	R/%	A_cc/%	F_score/%
Faster R⁃CNN^[19]	82.6	84.5	83.6	84.1
FCOS^[20]	84.7	88.1	86.1	87.4
DDN^[21]	83.8	88.5	87.1	87.5
ResNet^[22]	83.4	85.9	84.7	85.3
ConvNet^[23]	84.5	88.7	86.5	87.8
Multiscale⁃SSD^[15]	84.1	92.4	90.3	90.6
THMS^[24]	90.9	87.5	92.1	88.1
MFSA	92.4	95.8	94.7	95.1

表2 MFSA网络的消融实验结果

Table 2 Experimental Results of MFSA Network Ablation

数据增强模块	浅层特征抑制模块	Transformer 模块	F_score/%
			79.2
🗸			83.4
🗸	🗸		90.3
🗸		🗸	87.6
🗸	🗸	🗸	95.1

参考文献 24

1	姚睿，郎宪明，袁开欣，等.基于改进的Swin Transformer的输油管道微小缺陷检测模型［J］.管道技术与设备，2024（3）： 35⁃42.
	YAO R，LANG X M，YUAN K X，et al.Detection model of oil pipeline tiny defects based on improved Swin Transformer［J］.Pipeline Technology and Equipment，2024（3）：35⁃42.
2	王晓娟，杨帆，王错，等.面向超声导波检测的管道腐蚀建模及仿真实现［J］.仪器仪表学报，2023，44（5）：71⁃80.
	WANG X J，YANG F，WANG C，et al.Pipeline corrosion modeling and simulation for guided⁃waves⁃based inspection［J］. Chinese Journal of Scientific Instrument，2023，44（5）：71⁃80.
3	郭颖，杨理践，赵佰顺，等.长输油管道泄漏检测技术研究现状［J］.辽宁石油化工大学学报，2022，42（4）：25⁃31.
	GUO Y，YANG L J，ZHAO B S，et al.Research status of leakage detection technology for long⁃distance oil pipelines［J］. Journal of Liaoning Petrochemical University，2022，42（4）：25⁃31.
4	米亚鑫，郎宪明，韩富丞，等.基于低层特征融合多核卷积神经网络的管道缺陷漏磁图像识别方法［J］.当代化工，2023，52（3）：677⁃681.
	MI Y X，LANG X M，HAN F C，et al.Magnetic flux leakage image recognition method of pipeline defects based on low⁃level feature fusion multi⁃core convolution neural network［J］.Contemporary Chemical Industry，2023，52（3）：677⁃681.
5	杨理践，梁成壮，高松巍，等.管道漏磁内检测裂纹缺陷仿真与实验研究［J］.辽宁石油化工大学学报，2020，40（6）：59⁃63.
	YANG L J，LIANG C Z，GAO S W，et al.Simulation and experimental research on the crack defect of pipeline magnetic flux leakage inspection［J］.Journal of Liaoning Shihua University，2020，40（6）：59⁃63.
6	DONG H W，SONG K C，HE Y，et al.PGA⁃net：Pyramid feature fusion and global context attention network for automated surface defect detection［J］.IEEE Transactions on Industrial Informatics，2020，16（12）：7448⁃7458.
7	LIU T H，HE Z S.TAS2⁃net： Triple⁃attention semantic segmentation network for small surface defect detection［J］.IEEE Transactions on Instrumentation and Measurement，2022，71：1⁃12.
8	HSU W Y，CHEN Y S.Single image dehazing using wavelet⁃based haze⁃lines and denoising［J］.IEEE Access，2021，9： 104547⁃104559.
9	LI H L，SHI J H，LI L J，et al.Novel wavelet threshold denoising method to highlight the first break of noisy micro seismic recordings［J］.IEEE Transactions on Geoscience and Remote Sensing，2022，60：1⁃10.
10	NIU S L，LI B，WANG X G，et al.Defect image sample generation with GAN for improving defect recognition［J］.IEEE Transactions on Automation Science and Engineering，2020，17（3）：1611⁃1622.
11	GUO R Y，LIU H，XIE G，et al.Weld defect detection from imbalanced radiographic images based on contrast enhancement conditional generative adversarial network and transfer learning［J］.IEEE Sensors Journal，2021，21（9）：10844⁃10853.
12	JIANG J L，ZHU J L，BILAL M，et al.Masked swin transformer U⁃net for industrial anomaly detection［J］.IEEE Transactions on Industrial Informatics，2023，19（2）：2200⁃2209.
13	HAN F C，LANG X M.A fast magnetic flux leakage small defect detection network［J］.IEEE Transactions on Industrial Informatics，2023，19（12）：11941⁃11948.
14	LANG X M，HAN F C.MFL image recognition method of pipeline corrosion defects based on multilayer feature fusion multiscale GhostNet［J］.IEEE Transactions on Instrumentation and Measurement，2022，71：1⁃8.
15	YANG L J，WANG Z J，GAO S W.Pipeline magnetic flux leakage image detection algorithm based on multiscale SSD network［J］.IEEE Transactions on Industrial Informatics，2020，16（1）：501⁃509.
16	HAN F C，LANG X M，LIU M Y.An anchor⁃free pipeline MFL image detection method［J］.IEEE Transactions on Instrumentation and Measurement，2023，72：1⁃8.
17	ZENG N Y，WU P S，WANG Z D，et al.A small⁃sized object detection oriented multi⁃scale feature fusion approach with application to defect detection［J］.IEEE Transactions on Instrumentation and Measurement，2022，71：1⁃14.
18	LI D W，LI Y D，XIE Q，et al.Tiny defect detection in high⁃resolution aero⁃engine blade images via a coarse⁃to⁃fine framework［J］.IEEE Transactions on Instrumentation and Measurement，2021，70：1⁃12.
19	REN S Q，HE K M，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.
20	TIAN Z，SHEN C H，CHEN H，et al.FCOS：A simple and strong anchor⁃free object detector［J］.IEEE Transactions on Pattern Analysis and Machine Intelligence，2022，44（4）：1922⁃1933.
21	HE Y，SONG K C，MENG Q G，et al.An end⁃to⁃end steel surface defect detection approach via fusing multiple hierarchical features［J］.IEEE Transactions on Instrumentation and Measurement，2020，69（4）：1493⁃1504.
22	WANG W Y，MI C F，WU Z H， et al. A real⁃time steel surface defect detection approach with high accuracy［J］.IEEE Transactions on Instrumentation and Measurement，2022，71：1⁃10.
23	LIU Z，MAO H Z，WU C Y，et al.A ConvNet for the 2020s［C］//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）.Manhattan：IEEE，2022：11966⁃11976.
24	WANG C Y，BOCHKOVSKIY A，LIAO H Y M.YOLOv7： Trainable bag⁃of⁃freebies sets new state⁃of⁃the⁃art for real⁃time object detectors［EB/OL］.（2022⁃07⁃06）［2024⁃08⁃05］.https：//arxiv.org/abs/2207.02696.

[1]	李伟杰, 祁军, 潘斌. 基于通道注意力和时序改进多摄像头的鸟瞰视角目标检测[J]. 辽宁石油化工大学学报, 2024, 44(6): 89-96.
[2]	刘佳玮, 曹江涛, 姬晓飞. 基于Swin Transformer和YOLOv5的无纺布瑕疵检测[J]. 辽宁石油化工大学学报, 2024, 44(3): 80-88.
[3]	夏起磊, 罗林, 张垚. 基于注意力增强型编解码网络的化工过程故障诊断[J]. 辽宁石油化工大学学报, 2024, 44(2): 63-70.
[4]	赵旭峰, 刘琳琳, 曹宇, 叶成荫, 郭宗凯. 基于图像分块和多头注意力机制的气象识别研究[J]. 辽宁石油化工大学学报, 2024, 44(2): 83-90.
[5]	姜楠, 罗林, 王乔, 侯维. 结构优化深度网络的高压断路器机械故障诊断[J]. 辽宁石油化工大学学报, 2023, 43(3): 91-96.
[6]	汤永恒, 潘斌. 基于双向转换网络的域自适应单幅图像去雾方法[J]. 辽宁石油化工大学学报, 2022, 42(6): 78-83.
[7]	裴小邓，罗林，陈帅，王乔. 面向电力变压器油中溶解气体的卷积神经网络诊断方法[J]. 辽宁石油化工大学学报, 2020, 40(5): 79-85.
[8]	王竹筠，杨理践，高松巍. 基于Laplacian与多尺度数学形态学的漏磁图像边缘增强方法[J]. 辽宁石油化工大学学报, 2019, 39(5): 98-102.

基于浅层特征抑制的油气管道小缺陷检测网络

Detection Network for Small Defects in Oil and Gas Pipelines Based on Shallow Feature Suppression

HTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 24

相关文章 8

编辑推荐

Metrics