Abstract
Maintenance of railroad track safety is of utmost importance as derailment accidents cause significant loss to life and property. Inspection of railroad tracks and their components is necessary in order to ensure security and well-being of goods as well as humans. Fishplate is an essential component in the railroad track environment hence, periodic maintenance of fishplates is an imperative goal. In this paper, we propose a method for detection and segmentation of fishplate instances in high-altitude drone images (DI) for a closer-view and consequent inspection of fishplate instances. For this purpose, a novel two-stage Mask R-CNN-based framework termed as FishTwoMask R-CNN is proposed. A new fine-tuning strategy has been developed for the purpose of improving the detections in the second stage (Stage 2) which includes a training trick of modifying the loss weights for Stage 2 training. In the first stage (Stage 1), we detect fishplate instances, which are then cropped and fed as input to Stage 2, along with Stage 1 dataset. The Stage 2 network is then trained through a modified weighted loss and produces final detections for segmentation and further inspection. The”layers” hyper-parameter is assigned as “heads” for Stage 1 and updated to “4 + ” for Stage 2. Also, the critical analysis of Mask R-CNN hyper-parameters has been carried out during both the stages which has lead to an improved detection precision rate of 97% in Stage 2 as opposed to 47% in Stage 1. We evaluate our proposed approach on five different test image scenarios in order to view fishplate instance detection results. There has been statistical evaluation on out-of-distribution test images also in order to compute the metrics values. The comparative results have been evaluated using metrics of precision, recall, and F1-score on Mask R-CNN Stage 1 and Stage 2 along with Faster R-CNN and YOLOv5 methods. It is inferred that the proposed approach achieves appreciable metrics values and thus can be gathered suitable for fishplate instance segmentation in drone images.
Similar content being viewed by others
Data availability
The datasets generated during and/or analyzed during the current study are not publicly available due to privacy reason but are available from the corresponding author on reasonable request.
References
Abdulla W (2017) Mask r-cnn for object detection and instance segmentation on keras and tensorflow
Bharath B, Kanmani M (2017) Swarm intelligence based image fusion for thermal and visible images, in 2017 International Conference on Computation of Power, Energy Information and Commuincation (ICCPEIC), 043–048
Bhat S, Karegowda D, Noushad I (2021) Smart railway track monitoring system
Buggy SJ et al (2016) Railway track component condition monitoring using optical fibre Bragg grating sensors. Meas Sci Technol 27(5):055201
Chen Z et al (2022) Foreign object detection for railway ballastless trackbeds: A Semisupervised Learning Method. Measurement, 110757
Du C, Dutta S, Kurup P, Yu T, Wang X (2020) A review of railway infrastructure monitoring using fiber optic sensors. Sensors and Actuators A: Physical 303:111728. https://doi.org/10.1016/j.sna.2019.111728
Gao M, Wu H, Shen Y, Wang X, Zeng Y (2019) A peak detection algorithm adopting magnetic sensor signal for rail spike location in tamping machine. Adv Mech Eng 11(11):1687814019891570
Gavai G, Eldardiry H, Wu W, Xu B, Komatsu Y, Makino S (2019) Hybrid image-based defect detection for railroad maintenance. Electronic Imaging 2019(9):360–361
Güçlü E, Aydın İ, Akın E (2022) Measurement of railway sleepers spacing using mask R-CNN, in 2022 International Conference on Decision Aid Sciences and Applications (DASA), 1416–1420
Guo F, Qian Y, Rizos D, Suo Z, Chen X (2021) Automatic rail surface defects inspection based on Mask R-CNN. Transp Res Rec 2675(11):655–668
Guo F, Qian Y, Wu Y, Leng Z, Yu H (2021) Automatic railroad track components inspection using real-time instance segmentation. Computer-Aided Civil and Infrastructure Engineering 36(3):362–377
He K, Gkioxari G, Dollár P, Girshick R (2017) Mask r-cnn, in Proceedings of the IEEE international conference on computer vision, 2961–2969
Hodge VJ, O’Keefe S, Weeks M, Moulds A (2015) Wireless sensor networks for condition monitoring in the railway industry: A Survey. IEEE Transactions on Intelligent Transportation Systems 16(3):1088–1106. https://doi.org/10.1109/TITS.2014.2366512
Kafetzis D, Fourfouris I, Argyropoulos S, Koutsopoulos I (2020) UAV-assisted aerial survey of railways using deep learning, in 2020 International Conference on Unmanned Aircraft Systems (ICUAS), 1491–1500
Kanmani M, Narasimhan V (2019) An optimal weighted averaging fusion strategy for remotely sensed images. Multidimension Syst Signal Process 30(4):1911–1935
Kanmani M, Narasimhan V (2020) Optimal fusion aided face recognition from visible and thermal face images. Multimedia Tools and Applications 79(25):17859–17883
Liu J, Wang Z, Wu Y, Qin Y, Cao X, Huang Y (2020) An Improved Faster R-CNN for UAV-Based Catenary Support Device Inspection. Int J Software Eng Knowl Eng 30(07):941–959
Madheswari K, Venkateswaran N (2017) Swarm intelligence based optimisation in thermal image fusion using dual tree discrete wavelet transform. Quantitative Infrared Thermography Journal 14(1):24–43
Madheswari K, Venkateswaran N, Sowmiya V (2016) Visible and thermal image fusion using curvelet transform and brain storm optimization, in 2016 IEEE Region 10 Conference (TENCON), 2826–2829
Nayan MMR, Al Sufi S, Abedin AK, Ahamed R, Hossain MF (2020) An IoT based real-time railway fishplate monitoring system for early warning, in 2020 11th International Conference on Electrical and Computer Engineering (ICECE), 310–313
Rampriya RS, Suganya R, Nathan S, Perumal PS (2022) A comparative assessment of deep neural network models for detecting obstacles in the real time aerial railway track images. Applied Artificial Intelligence, 1–33
Ravichandran A, Raja A, Kanmani M (2017) Entropy optimized image fusion: Using particle swarm technology and discrete wavelet transform, in 2017 international conference on computation of power, energy information and commuincation (ICCPEIC), 068–074
Saini A, Agarwal A, Singh D (2020) Feature-based template matching for joggled fishplate detection in railroad track with drone images, in IGARSS 2020–2020 IEEE International Geoscience and Remote Sensing Symposium, 2237–2240
Saini A, Kishore KG, Sriram KSS, Singh D, Singh KP (2022) Machine learning approach for detection of track assets for railroad health monitoring with drone images, in IGARSS 2022–2022 IEEE International Geoscience and Remote Sensing Symposium, 4891–4894
Saini A, Singh D (2018) Development of computer vision based robust approach for joggled fish plate detection in drone images, in 2018 9th International Symposium on Signal, Image, Video and Communications (ISIVC), 33–38
Saini A, Singh D (2021) DroneRTEF: development of a novel adaptive framework for railroad track extraction in drone images. Pattern Anal Appl 24(4):1549–1568
Singh P, Garg RD (2014) Classification of high resolution satellite images using spatial constraints-based fuzzy clustering. J Appl Remote Sens 8(1):083526
Singh PP, Garg RD (2015) Fixed point ICA based approach for maximizing the non-Gaussianity in remote sensing image classification. Journal of the Indian Society of Remote Sensing 43(4):851–858
Sumit SS, Watada J, Roy A, Rambli DRA (2020) In object detection deep learning methods, YOLO shows supremum to Mask R-CNN. J Phys: Conf Ser 1529(4):042086
Szeliski R (2010) Computer vision: algorithms and applications. Springer Science & Business Media
Wu Y, Meng F, Qin Y, Qian Y, Xu F, Jia L (2023) UAV imagery based potential safety hazard evaluation for high-speed railroad using Real-time instance segmentation. Adv Eng Inform 55:101819
Yilmazer M, Karakose M (2022) “Mask R-CNN architecture based railway fastener fault detection approach”. International Conference on Decision Aid Sciences and Applications (DASA) 2022:1363–1366
Zheng D et al (2021) A defect detection method for rail surface and fasteners based on deep convolutional neural network. Computational Intelligence and Neuroscience
Acknowledgements
The authors would like to thank RailTel, India for supporting this work. The authors would also like to extend their thanks to British Council and IIT Roorkee for granting Newton Bhabha Fund under which a part of this research work has been carried out at The University of Sheffield, Sheffield, United Kingdom.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors would like to state that there is no conflict of interest amongst any of the authors.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Saini, A., Singh, D. & Alvarez, M. FishTwoMask R-CNN: Two-stage Mask R-CNN approach for detection of fishplates in high-altitude railroad track drone images. Multimed Tools Appl 83, 10367–10392 (2024). https://doi.org/10.1007/s11042-023-15924-7
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11042-023-15924-7