An ensemble method for automatic real-time detection, evaluation and position of exposed subsea pipelines based on 3D real-time sonar system

At present, traditional subsea pipeline structural health monitoring (SHM) uses the sonar equipment, which needs to post-process the obtained data. Then the sonar images can be obtained. Moreover, traditional SHM needs staff to interpret the results of the sonar images. Such results can be subject to manual interference, and monitoring efficiency and accuracy cannot be guaranteed either. In view of the above problems, this paper proposed an ensemble method for real-time automatic monitoring, evaluation and positioning of exposed subsea pipelines based on 3D real-time sonar system. The method first used the YOLO V5 algorithm to realize the automatic identification of exposed pipelines in 3D real-time sonar images. It simplified the target detector in the YOLO V5 algorithm, further enhancing the real-time detection performance of the algorithm. A primary evaluation of the exposure of the pipelines was then performed by the region grow algorithm. The starting point of the algorithm was the center point coordinates of the detection box identified by YOLO V5. Finally, the positioning of the exposed pipeline was completed through the spatial position mapping relationship between the subsea pipeline, the Remote Operated Vehicle (ROV) and the tracking ship. The ensemble method proposed in this study can operate automated real-time monitoring of exposed subsea pipelines. Besides, it can also evaluate exposure degree and positioning of the pipelines. This can effectively avoid manual interference on the interpretation of sonar images, and guarantee the monitoring efficiency and accuracy of exposed subsea pipelines. The feasibility of this method was verified by experiments. It is hoped that the ensemble method proposed in this paper can provide a technical reference for the practical application of such tasks in the future.

Automated summary

This study proposes an ensemble method based on a 3D real-time sonar system for automatic monitoring, evaluation, and positioning of exposed subsea pipelines. The method uses the YOLO V5 algorithm for automatic pipeline identification and the region grow algorithm for initial exposure evaluation. The pipeline positioning is achieved through the spatial position mapping relationship.