|
Article information
2023 , Volume 28, ¹ 3, p.182-194
Smagin S.I., Smagin A.S.
Automated search for damage to underwater mesh fences
Purpose. The aim of the work is to develop and study image processing algorithms designed to solve the problems of visual monitoring of the integrity and contamination of the mesh enclosing structures for underwater fish farms. Methodology. The theoretical basis for the research relies on the methods and approaches used in the theory of pattern recognition, such as the analysis of the frequency characteristics of images using the Retinex transformation, the method of adaptive binarization of Otsu. Computer vision and machine learning technologies were used to develop algorithms for analyzing parametric contours and filtering masks of objects. Findings. 1. New computer vision algorithms have been developed to assess the condition of mesh fences by their binary masks. This eliminates the processing of unnecessary data in the image, reduces time and increases the accuracy of damage detection. 2. A new modification of the convolution for neural network architecture has been developed for the problem of semantic segmentation of mesh fencing, the computational complexity of which is lower than that of the basic architectures described in the scientific literature. The proposed approach is based on the use of learnable regularization (residual blocks), which allows obtaining an object mask of higher quality than those obtained by classical computer vision. 3. A software package has been developed to automate the visual monitoring of underwater mesh fences, generating and using a binary fence mask to assess its condition. Originality/value. The algorithms developed in the dissertation showed sufficiently high performance and significantly higher accuracy (from 88 to 100 %, depending on the algorithm used) for determining damage to underwater mesh fences than those proposed earlier in the works of other authors.
Keywords: automation, computer vision, machine learning, underwater vehicle, software package
doi: 10.25743/ICT.2023.28.3.011
Author(s):
Smagin Sergey Ivanovich
Dr. , Correspondent member of RAS, Professor
Position: Director
Office: Computer Center FEB RAS
Address: 680000, Russia, Khabarovsk
Phone Office: (4212) 22 72 67
E-mail: smagin@ccfebras.ru
SPIN-code: 2419-4990Smagin Alexey Sergeevich
Position: Junior Research Scientist
Office: Mining Institute of Far-Eastern branch of Russian Academy of Science
Address: 680000, Russia, Khabarovsk, Ussury Boulevard,5
E-mail: smaginkhv@gmail.com
References:
1. Sergeev L., Kuzin V., Kharin A., Mnatsakanyan R., Mnatsakanyan A. Ekonomika rybnogo khozyaystva. Tsifrovizatsiya upravleniya. Uchebnik dlya vuzov [Economics of fisheries. Digitalization of management. Textbook for universities]. Moscow: Izdatel’stvo Yurayt; 2023: 318. (In Russ.)
2. Colbourne D.B. Another perspective on challenges in open ocean aquaculture development. IEEE Journal of Oceanic Engineering. 2005: 30(1):4–11.
3. Huikai W., Junge Z., Kaiqi H., Kongming L., Yizhou Y. FastFCN: rethinking dilated convolution in the backbone for semantic segmentation. 2019: 1–15. Available at: https://arxiv.org/pdf/1903.11816.pdf.
4. Garcia R., Nicosevici T., Cufi X. On the way to solve lighting problems in underwater imaging.OCEANS’02 MTS/IEEE. IEEE; 2002; (2):1018–1024.
5. Wang J., Lu K., Xue J., He N., Shao L. Single image dehazing based on the physical model and MSRCR algorithm. IEEE Transactions on Circuits and Systems for Video Technology. 2017;28(9):2190–2199.
6. Smagin A.S., Dubrovin K.N. On computer vision algorithms for searching breaks in meshed fencing constructions. Computational Technologies. 2019; 24(6):118–125.DOI:10.25743/ICT.2019.24.6.014. (In Russ.)
7. Israfilov H.S. Research of methods for binarization of images. Herald of Science and Education.2017; 2(6):43–50.
8. Guo Y., Liu Y., Georgiou T., Lew M.S. A review of semantic segmentation using deep neural networks. International Journal of Multimedia Information Retrieval. 2018; (7):87–93.
9. Zhou D., Fang J., Song X., Guan C., Yin J., Dai Y., Yang R. Iou loss for 2d/3d object detection. 2019 International Conference on 3D Vision (3DV). 2019: 85–94.
10. Kachalin S.V. Increasing the learning stability of large neural networks by supplementing small training samples of parent examples with synthesized biometric descendant examples. Materialy Nauchno-Tekhnicheskoy Konferentsii Klastera Penzenskikh Predpriyatiy, Obespechivayushchikh Bezopasnost’ Informatsionnykh Tekhnologiy [Proceedings of the Scientific and Technical Conference of the
Cluster of Penza Enterprises Providing Information Technology Security]. Penza; 2014; (9):32–35.(In Russ.)
11. Yaeger L., Lyon R., Webb B. Effective training of a neural network character classifier for word recognition. NIPS. 1996. Available at: https://www.semanticscholar.org/paper/EffectiveTraining-of-a-Neural-Network-Character-Yaeger Lyon/437fce6c281031a9dc69db9c54027b531dcbeecc.
12. Ciresan D., Meier U., Gambardella L., Schmidhuber J. Deep big simple neural nets excel on handwritten digit recognition. Neural Computation. 2010; 22(12). Available at: https://arxiv.org/abs/1003.0358.
13. Simard P., Steinkraus D., Platt J. Best practices for convolutional neural networks applied to visual document analysis. International Conference on Document Analysis and Recognition. IEEE;2003. DOI:10.1109/ICDAR.2003.1227801. Available at: https://ieeexplore.ieee.org/document/ 1227801.
14. Chakraverty S., Sahoo D.M., Mahato N.R. McCulloch – Pitts neural network model. Concepts of Soft Computing. Springer; 2019: 167–173. Available at: https://link.springer.com/chapter/10.1007/978-981-13-7430-2_11.
15. Ronneberger O., Fischer P., Brox T. U-net: convolutional networks for biomedical image segmentation. Medical Image Computing and Computer-Assisted Intervention — MICCAI 2015: 18th International Conference. Springer International Publishing; 2015; 3(18):234–241.
16. Badrinarayanan V., Kendall A., Cipolla R. Segnet: a deep convolutional encoder-decoder architecture for image segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence.2017; 39(12):2481–2495.
17. He K., Gkioxari G., Dollar P., Girshick R. Mask R-CNN. 2018; 1–12. Available at: https://arxiv.org/pdf/1703.06870.pdf.
18. Wang Q., Ma Y., Zhao K., Tian Y. A comprehensive survey of loss functions in machine learning.Annals of Data Science. 2022; (9):187–212. Available at: https://link.springer.com/article/10.1007/s40745-020-00253-5.
19. Fathi A. Semantic instance segmentation via deep metric learning. 2019. Available at: https://arxiv.org/pdf/1703.10277.pdf.
20. Rosebrock A. Intersection over Union (IoU) for object detection. 2016. Available at: https://pyimagesearch.com/2016/11/07/intersection-over-union-iou-for-object-detection.
21. Papandreou G. Weakly-and semi-supervised learning of a deep convolutional network for semantic image segmentation. Proceedings of the IEEE International Conference on Computer Vision.2015: 1742–1750.
22. Bradski G. The OpenCV library. Dr. Dobb’s Journal: Software Tools for the Professional Programmer.2000; 25(11):120–123.
23. Smagin A.S. Algoritmy komp’yuternogo zreniya dlya otsenki sostoyaniya podvodnykh setchatykh ograzhdeniy [Computer vision algorithms for assessing the condition of underwater mesh fences]. PhD Thesis: 2.3.1. 2022: 145. (In Russ.)
Bibliography link:
Smagin S.I., Smagin A.S. Automated search for damage to underwater mesh fences // Computational technologies. 2023. V. 28. ¹ 3. P. 182-194
|
