Automated Pipe Defect Identification in Underwater Robot Imagery with Deep Learning

Abstract

Underwater pipeline inspection plays a vital role in the proactive maintenance and management of critical marine infrastructure and subaquatic systems. However, the inspection of underwater pipelines presents a challenge due to factors such as light scattering, absorption, restricted visibility, and ambient noise. The advancement of deep learning has introduced powerful techniques for processing large amounts of unstructured and imperfect data collected from underwater environments. This study evaluated the efficacy of the You Only Look Once (YOLO) algorithm, a real-time object detection and localization model based on convolutional neural networks, in identifying and classifying various types of pipeline defects in underwater settings. YOLOv8, the latest evolution in the YOLO family, integrates advanced capabilities, such as anchor-free detection, a cross-stage partial network backbone for efficient feature extraction, and a feature pyramid network+ path aggregation network neck for robust multi-scale object detection, which make it particularly well-suited for complex underwater environments. Due to the lack of suitable open-access datasets for underwater pipeline defects, a custom dataset was captured using a remotely operated vehicle in a controlled environment. This application has the following assets available for use. Extensive experimentation demonstrated that YOLOv8 X-Large consistently outperformed other models in terms of pipe defect detection and classification and achieved a strong balance between precision and recall in identifying pipeline cracks, rust, corners, defective welds, flanges, tapes, and holes. This research establishes the baseline performance of YOLOv8 for underwater defect detection and showcases its potential to enhance the reliability and efficiency of pipeline inspection tasks in challenging underwater environments.

Keywords

YOLO8 / Underwater robot / Object detection / Underwater pipelines / Remotely operated vehicle / Deep learning

References

[1] AboahA, WangB, BagciU, Adu-GyamfiY. Real-time multiclass helmet violation detection using few-shot data sampling technique and YOLOv8. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023 Piscataway IEEE 5349-5357

[2] AdarshP, RathiP, KumarM. YOLO v3-Tiny: Object detection and recognition using one stage improved model. Proceedings of the 2020 6th International Conference on Advanced Computing and Communication Systems (ICACCS), 2020 Piscataway IEEE 687-694

[3] Al MuksitA, HasanF, EmonMFHB, HaqueMR, AnwaryAR, ShatabdaS. YOLO-fish: A robust fish detection model to detect fish in realistic underwater environment. Ecological Informatics, 2022, 72: 101847

[4] AsyrafMS, IsaIS, MarzukiMIF, SulaimanSN, HungCC. CNN-based YOLO v3 comparison for underwater object detection. Journal of Electrical and Electronic Systems Research, 2021, 18: 30-37

[5] AvsarE, FeekingsJP, KragLA. Estimating catch rates in real time: Development of a deep learning based Nephrops (Nephrops norvegicus) counter for demersal trawl fisheries. Frontiers in Marine Science, 2023, 10: 1129852

[6] BurgueraA, Bonin-FontF. Advances in autonomous underwater robotics based on machine learning. Journal of Marine Science and Engineering, 2022, 10(10): 1481

[7] CaoC, YuY, XieY, SunC. An efficient approach for gastric polyps detection based on improved SSD. Proceedings of the 2021 China Automation Congress (CAC), 2021 Piscataway IEEE 852-857

[8] ChenY, LiQ, LuD, KouL, KeW, BaiY, WangZ. A novel underwater image enhancement using optimal composite backbone network. Biomimetics, 2023, 8(3): 275

[9] ChengC. Real-time mask detection based on SSDMobileNetV2. In Proceedings of the 2022 IEEE 5th International Conference on Automation, Electronics and Electrical Engineering (AUTEEE), 2022 Piscataway IEEE 761-767

[10] ChiY, ZhangC. Underwater image enhancement methods using biovision and type-II fuzzy set. Journal of Marine Science and Engineering, 2024, 12(11): 2080

[11] DuP, SongX. Lightweight target detection: An improved YOLOv8 for small target defect detection on printed circuit boards. Proceedings of the 2024 International Conference on Generative Artificial Intelligence and Information Security (GAIIS), 2024 New York ACM 329-334

[12] FayazS, ParahSA, QureshiGJ. Underwater object detection: Architectures and algorithms—A comprehensive review. Multimedia Tools and Applications, 2022, 81(15): 20871-20916

[13] FuC, LiuR, FanX, ChenP, FuH, YuanW, ZhuM, LuoZ. Rethinking general underwater object detection: Datasets, challenges, and solutions. Neurocomputing, 2023, 517: 243-256

[14] GaoY, LiuW, ChuiH-C, ChenX. Large span sizes and irregular shapes target detection methods using variable convolutionimproved YOLOv8. Sensors, 2024, 24(8): 2560

[15] GašparovićB, LergaJ, MaušaG, Ivašić-KosM. Deep learning approach for objects detection in underwater pipeline images. Applied Artificial Intelligence, 2022, 36(1): 2146853

[16] GirshickR, DonahueJ, DarrellT, MalikJ. Rich feature hierarchies for accurate object detection and semantic segmentation. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2014 Piscataway IEEE 580-587

[17] HanM, LyuZ, QiuT, XuM. A review on intelligence dehazing and color restoration for underwater images. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2020, 50(5): 1820-1832

[18] HeK, ZhangX, RenS, SunJ. Spatial pyramid pooling in deep convolutional networks for visual recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(9): 1904-1916

[19] JinZZ, ZhengYF. Research on application of improved YOLO V3 algorithm in road target detection. Journal of Physics: Conference Series, 2020, 1654: 012060

[20] JocherG, ChaurasiaA, StokenA, BorovecJ NanoCode012. ultralytics/yolov5: v7.0-YOLOv5 SOTA realtime instance segmentation. Zenodo, 2022

[21] KarimanziraD, RenkewitzH, SheaD, AlbiezJ. Object detection in sonar images. Electronics, 2020, 9(7): 1180

[22] KimJH, KimN, WonCS. High-speed drone detection based on YOLO-V8. Proceedings of the 2023 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2023 Piscataway IEEE 1-2

[23] KimN, KimJH, WonCS. FAFD: Fast and accurate face detector. Electronics, 2022, 11(6): 875

[24] LiC, LiL, JiangH, WengK, GengY, LiL, KeZ, LiQ, ChengM, NieW, LiY, ZhangB, LiangY, ZhouL, XuX, ChuX, WeiX, WeiX YOLOv6: A single-stage object detection framework for industrial applications, 2022

[25] LiH, GuZ, HeD, WangX, HuangJ, MoY, LiP, HuangZ, WuF. A lightweight improved YOLOv5s model and its deployment for detecting pitaya fruits in daytime and nighttime light-supplement environments. Computers and Electronics in Agriculture, 2024, 220: 108914

[26] LiuK, SunQ, SunD, PengL, YangM, WangN. Underwater target detection based on improved YOLOv7. Journal of Marine Science and Engineering, 2023, 11(3): 677

[27] LiuW, AnguelovD, ErhanD, SzegedyC, ReedS, FuCY, BergAC. SSD: Single shot multiBox detector. Computer Vision — ECCV 2016: 14th European Conference, 2016 Berlin Springer Part I 21-37

[28] MengF, LiJ, ZhangY, QiS, TangY. Transforming unmanned pineapple picking with spatio-temporal convolutional neural networks. Computers and Electronics in Agriculture, 2023, 214: 108298

[29] NguyenQH, LyHB, HoLS, Al-AnsariN, LeHV, TranVQ, PrakashI, PhamBT. Influence of data splitting on performance of machine learning models in prediction of shear strength of soil. Mathematical Problems in Engineering, 2021 4832864

[30] OrinaitėU, KaraliūtėV, PalM, RagulskisM. Detecting underwater concrete cracks with machine learning: A clear vision of a murky problem. Applied Sciences, 2023, 13(12): 7335

[31] Park CW, Eom IK (2024) Underwater image enhancement using adaptive standardization and normalization networks. Engineering Applications of Artificial Intelligence 127(Part A): 107445. https://doi.org/10.1016/j.engappai.2023.107445

[32] PavaniD, ReddyANN, SawN, PrasadS, NaikSM. Octacleaner: Underwater trash detection through YOLO. Proceedings of the 2023 3rd International Conference on Mobile Networks and Wireless Communications (ICMNWC), 2023 Piscataway IEEE 1-6

[33] QYSEA FIFISH V6 plus official website, 2022

[34] RedmonJ, DivvalaS, GirshickR, FarhadiA. You only look once: Unified, real-time object detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016 Piscataway IEEE 779-788

[35] RahmanS, RonyJH, UddinJ, SamadMA. Real-time obstacle detection with YOLOv8 in a WSN using UAV aerial photography. Journal of Imaging, 2023, 9(10): 216

[36] RazaK, HongS. Fast and accurate fish detection design with improved YOLO-v3 model and transfer learning. International Journal of Advanced Computer Science and Applications, 2020, 11(2): 7-16

[37] RedmonJ, FarhadiA YOLOv3: An incremental improvement, 2018

[38] RenS, HeK, GirshickR, SunJ. Faster R-CNN: Towards realtime object detection with region proposal networks. Advances in Neural Information Processing Systems 28 (NeurIPS), 2015

[39] SelcukB, SerifT. A comparison of YOLOv5 and YOLOv8 in the context of mobile UI detection. Proceedings of the International Conference on Mobile Web and Intelligent Information Systems, 2023 161-174

[40] ShankarR, MuthulakshmiM. Comparing YOLOv3, YOLOv5 & YOLOv7 architectures for underwater marine creatures detection. Proceedings of the 2023 International Conference on Computational Intelligence and Knowledge Economy (ICCIKE), 2023 Piscataway IEEE 25-30

[41] ShenL, TaoH, NiY, WangY, StojanovicV. Improved YOLOv3 model with feature map cropping for multi-scale road object detection. Measurement Science and Technology, 2023, 34(4): 045406

[42] SoormaMS, ChaudharyA, SonaliS, PalS, UpadhyayDK. Underwater image processing with normalized AttUNet. 2023 International Conference on Computer, Electronics & Electrical Engineering & their Applications (IC2E3), 2023 Piscataway IEEE 1-5

[43] VidhyaSK, DeepthiPS. A comprehensive analysis of underwater image processing based on deep learning techniques. Proceedings of the 2023 International Conference on Control, Communication and Computing (ICCC), 2023 Piscataway IEEE 1-6

[44] WangT, LiY, ZhaiY, WangW, HuangR. A sewer pipeline defect detection method based on improved YOLOv5. Processes, 2023, 118: 2508

[45] XuS, ZhangM, SongW, MeiH, HeQ, LiottaA. A systematic review and analysis of deep learning-based underwater object detection. Neurocomputing, 2023, 527: 204-232

[46] YaseenM What is YOLOv9: An in-depth exploration of the internal features of the next-generation object detector, 2024

[47] ZhangH, ZhangS, WangY, LiuY, YangY, ZhouT, BianH. Subsea pipeline leak inspection by autonomous underwater vehicle. Applied Ocean Research, 2021, 107: 102321

[48] ZhangH, DaiC, ChenC, ZhaoZ, LinM. One stage multiscale efficient network for underwater target detection. Review of Scientific Instruments, 2024, 956: 065108

[49] ZhangJ, LiuX, ZhangX, XiZ, WangS. Automatic detection method of sewer pipe defects using deep learning techniques. Applied Sciences, 2023, 137: 4589

[50] ZhangL, LinL, LiangX, HeK. Is faster R-CNN doing well for pedestrian detection? Computer Vision—ECCV 2016: 14th European Conference. ECCV 2016. Lecture Notes in Computer Science Part II: 443-457, 2016

[51] ZhangY, NiQ. A novel weld-seam defect detection algorithm based on the S-YOLO model. Axioms, 2023, 127: 697

[52] ZhouH, KongM, YuanH, PanY, WangX, ChenR, LuW, WangR, YangQ. Real-time underwater object detection technology for complex underwater environments based on deep learning. Ecological Informatics, 2024, 82: 102680

[53] ZhongJ, GaoC, TianY, ZhangM. Research on the influence of hydrodynamic analysis to dynamic modeling of underwater manipulator. Proceedings of the 2023 IEEE 6th Information Technology, Networking, Electronic and Automation Control Conference (ITNEC), 2023, 6: 982-986

[54] ZhaoS, ZhengJ, SunS, ZhangL. An improved YOLO algorithm for fast and accurate underwater object detection. Symmetry, 2022, 148: 1669

RIGHTS & PERMISSIONS

Harbin Engineering University and Springer-Verlag GmbH Germany, part of Springer Nature