Object Recognition in Underwater Environments Using AI Computer Vision Techniques
Abstract
Towards this end, we make use of the Semantic Segmentation of Underwater Imagery (SUIM) dataset, which consists of over 1.5k densely annotated photos from eight different item categories that have ground-truth examples. Vertebrate fish, invertebrate reefs, aquatic plants, robots, human divers (like me!), and even the seafloor are among the more than 2,000 categories. This dataset reminds us of organized synthesis by gathering data from multiple ocean expeditions and collaborative experiments with both humans (unmanned)and robots. The same authors' team published a more current work that included a thorough performance benchmarking using cutting-edge global representations that are easily downloadable as open-source code. When it comes to underwater Inspection, Maintenance, and Repair (IMR) duties, the assessed methods facilitate the use of Autonomous Underwater Vehicles (AUVs) for autonomous interventions. A selection of test objects was made that is indicative of the types of applications that use IMR and whose shapes are usually known in advance. As a result, in realistic settings, CAD models produce virtual representations of these things when noise is added, and resolution is decreased. We validated our approach through extensive testing on both simulated scans and real data obtained using an AUV combined with an in-house rapid laser scanning sensor. Additionally, testing was done underwater in areas where shifting terrain caused by an unstable bed may have altered the contour of items being followed. To show how it broadens the scope, the research goes deeper into evaluating the performance of cutting-edge semantic segmentation algorithms using recognized measures. Finally, we present a fully convolutional encoder-decoder model which is tailored for competitive performance and computational efficiency. The model achieved 88% accuracy which is very high as far as underwater image segmentation goes. This study shows how the model could be put to practical use in various tasks from visual serving, saliency prediction and complex scene understanding. Importantly, the ESRGAN utilization improves images quality that enriches the soil on which our model succeeds. It lays a strong foundation for forthcoming research in the field of underwater robot vision through formulation, modeling, and introduction to benchmark dataset.
References
2. Banno, K.; Kaland, H.; Crescitelli, A.M.; Tuene, S.A.; Aas, G.H.; Gansel, L.C. A novel approach for wild fish monitoring at aquaculture sites: Wild fish presence analysis using computer vision. Aquac. Environ. Interact. 2022, 14, 97–112.[CrossRef]
3. Saleh, A.; Sheaves, M.; Rahimi Azghadi, M. Computer vision and deep learning for fish classification in underwater habitats: A survey. Fish Fish. 2022, 23, 977–999. [CrossRef]
4. Ditria, E.M.; Sievers, M.; Lopez-Marcano, S.; Jinks, E.L.; Connolly, R.M. Deep learning for automated analysis of fish abundance: The benefits of training across multiple habitats. Environ. Monit. Assess. 2020, 192, 698. [CrossRef] [PubMed]
5. Ditria, E.M.; Lopez-Marcano, S.; Sievers, M.; Jinks, E.L.; Brown, C.J.; Connolly, R.M. Automating the Analysis of Fish Abundance Using Object Detection: Optimizing Animal Ecology With Deep Learning. Front. Mar. Sci. 2020, 7. [CrossRef]
6. Shafait, F.; Mian, A.; Shortis, M.; Ghanem, B.; Culverhouse, P.F.; Edgington, D.; Cline, D.; Ravanbakhsh, M.; Seager, J.; Harvey, E.S. Fish identification from videos captured in uncontrolled underwater environments. ICES J. Mar. Sci. 2016, 73, 2737–2746. [CrossRef]
7. Noda, J.J.; Travieso, C.M.; Sánchez-Rodríguez, D. Automatic Taxonomic Classification of Fish Based on Their Acoustic Signals. Appl. Sci. 2016, 6, 443. [CrossRef]
8. Helminen, J.; Linnansaari, T. Object and behavior differentiation for improved automated counts of migrating river fish using imaging sonar data. Fish. Res. 2021, 237, 105883. [CrossRef]
9. Saberioon, M.; Gholizadeh, A.; Cisar, P.; Pautsina, A.; Urban, J. Application of machine vision systems in aquaculture with emphasis on fish: state-of-the-art and key issues. Rev. Aquac. 2017, 9, 369–387. [CrossRef]
10. Salman, A.; Siddiqui, S.A.; Shafait, F.; Mian, A.; Shortis, M.R.; Khurshid, K.; Ulges, A.; Schwanecke, U. Automatic fish detection in underwater videos by a deep neural network-based hybrid motion learning system. ICES J. Mar. Sci. 2020, 77, 1295–1307.
11. Kim D, Lee D, Myung H, Choi H-TJISR. Artificial landmark-based underwater localization for AUVs using weighted template matching. 2014;7:175-84. https://doi.org/10.1007/s11370-014- 0153-y.
12. Chuang M-C, Hwang J-N, Williams KJIToIP. A feature learning and object recognition framework for underwater fish images. 2016;25(4):1862-72. https://doi.org/10.48550/arXiv.1603.01696.
13. Alaba, S. Y., Nabi, M. M., Shah, C., Prior, J., Campbell, M. D., Wallace, F., ... & Moorhead, R. (2022). Class-aware fish species recognition using deep learning for an imbalanced dataset. Sensors, 22(21), 8268. https://doi.org/10.3390/s22218268.
14. Villon S, Chaumont M, Subsol G, Villéger S, Claverie T, Mouillot D, editors. Coral reef fish detection and recognition in underwater videos by supervised machine learning: Comparison between Deep Learning and HOG+ SVM methods. Advanced Concepts for Intelligent Vision Systems: 17th International Conference, ACIVS 2016, Lecce, Italy, October 24-27, 2016, Proceedings 17; 2016: Springer. https://doi.org/10.1007/978-3-319-48680-2_15.
15. A. K. Gupta, A. Seal, M. Prasad, and P. J. E. Khanna, "Salient object detection techniques in computer vision—A survey," vol. 22, no. 10, p. 1174, 2020.
16. N. Chen, W. Liu, R. Bai, and A. J. A. I. R. Chen, "Application of computational intelligence technologies in emergency management: a literature review," vol. 52, pp. 2131-2168, 2019.
17. H. Qin, X. Li, J. Liang, Y. Peng, and C. J. N. Zhang, "Deep Fish: Accurate underwater live fish recognition with a deep architecture," vol. 187, pp. 49-58, 2016.
18. H. Huang, H. Zhou, X. Yang, L. Zhang, L. Qi, and A.-Y. J. N. Zang, "Faster R-CNN for marine organisms’ detection and recognition using data augmentation," vol. 337, pp. 372-384, 2019.
19. LeCun Y, Bengio Y, Hinton GJn. Deep learning. 2015;521(7553):436-44. http://dx.doi.org/10.1038/nature14539.
20. Aruna, S. K., Deepa, N., & Devi, T. (2023, May). Underwater Fish Identification in Real-Time using Convolutional Neural Network. In 2023 7th International Conference on Intelligent Computing and Control Systems (ICICCS) (pp. 586-591). IEEE. https://doi.org/10.1109/ICICCS56967.2023.10142531.
21. Zhao, D., Yang, B., Dou, Y., & Guo, X. (2022, November). Underwater fish detection in sonar image based on an improved Faster RCNN. In 2022 9th International Forum on Electrical Engineering and Automation (IFEEA) (pp. 358-363). IEEE. https://doi.org/10.1109/IFEEA57288.2022.10038226.
22. Han, G., Huang, S., Ma, J., He, Y., & Chang, S. F. (2022, June). Meta faster r-cnn: Towards accurate few-shot object detection with attentive feature alignment. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 36, No. 1, pp. 780-789). https://doi.org/10.1609/aaai.v36i1.19959.
23. Yang, H., Liu, P., Hu, Y., & Fu, J. (2021). Research on underwater object recognition based on YOLOv3. Microsystem Technologies, 27, 1837-1844. https://doi.org/10.1007/s00542-019-04694-8.
24. Shen, L., Tao, H., Ni, Y., Wang, Y., & Stojanovic, V. (2023). Improved YOLOv3 model with feature map cropping for multiscale road object detection. Measurement Science and Technology, 34(4), 045406. http://dx.doi.org/10.1088/1361-6501/acb075.
25. Bosse, S., & Kasundra, P. (2022). Robust Underwater Image Classification Using Image Segmentation, CNN, and Dynamic ROI Approximation. Engineering Proceedings, 27(1), 82. https://doi.org/10.3390/ecsa-9-13218.
26. Chen, Z., Wang, Y., Tian, W., Liu, J., Zhou, Y., & Shen, J. (2022). Underwater sonar image segmentation combining pixel-level and region-level information. Computers and Electrical Engineering, 100, 107853. https://doi.org/10.1016/j.compeleceng.2022.107853.
27. Wang, J., He, X., Shao, F., Lu, G., Hu, R., & Jiang, Q. (2022). Semantic segmentation method of underwater images based on encoder-decoder architecture. Plos one, 17(8), e0272666. https://doi.org/10.1371/journal.pone.0272666.
28. Liu Z, Tong L, Chen L, Zhou F, Jiang Z, Zhang Q, et al. Canet: Context aware network for brain glioma segmentation. 2021;40(7):1763-77. https://doi.org/10.1109/tmi.2021.3065918.
29. Alavianmehr, M. A., Helfroush, M. S., Danyali, H., & Tashk, A. (2023). Butterfly network: a convolutional neural network with a new architecture for multi-scale semantic segmentation of pedestrians. Journal of real-time image processing.
30. Dakhil, R. A., & Khayeat, A. R. H. (2022). Review On Deep Learning Technique For Underwater Object Detection. arXiv preprint arXiv:2209.10151. https://doi.org/10.48550/arXiv.2209.10151.
