Zero-Shot Pipeline Detection for Sub-Bottom Profiler Data Based on Imaging Principles

With the increasing number of underwater pipeline investigation activities, the research on automatic pipeline detection is of great significance. At this stage, object detection algorithms based on Deep Learning (DL) are widely used due to their abilities to deal with various complex scenarios. How...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Gen Zheng, Jianhu Zhao, Shaobo Li, Jie Feng
Formato: article
Lenguaje:EN
Publicado: MDPI AG 2021
Materias:
Q
Acceso en línea:https://doaj.org/article/a9e6144e15a04f389958b3deb25ad5ce
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:a9e6144e15a04f389958b3deb25ad5ce
record_format dspace
spelling oai:doaj.org-article:a9e6144e15a04f389958b3deb25ad5ce2021-11-11T18:55:34ZZero-Shot Pipeline Detection for Sub-Bottom Profiler Data Based on Imaging Principles10.3390/rs132144012072-4292https://doaj.org/article/a9e6144e15a04f389958b3deb25ad5ce2021-11-01T00:00:00Zhttps://www.mdpi.com/2072-4292/13/21/4401https://doaj.org/toc/2072-4292With the increasing number of underwater pipeline investigation activities, the research on automatic pipeline detection is of great significance. At this stage, object detection algorithms based on Deep Learning (DL) are widely used due to their abilities to deal with various complex scenarios. However, DL algorithms require massive representative samples, which are difficult to obtain for pipeline detection with sub-bottom profiler (SBP) data. In this paper, a zero-shot pipeline detection method is proposed. First, an efficient sample synthesis method based on SBP imaging principles is proposed to generate samples. Then, the generated samples are used to train the YOLOv5s network and a pipeline detection strategy is developed to meet the real-time requirements. Finally, the trained model is tested with the measured data. In the experiment, the trained model achieved a mAP@0.5 of 0.962, and the mean deviation of the predicted pipeline position is 0.23 pixels with a standard deviation of 1.94 pixels in the horizontal direction and 0.34 pixels with a standard deviation of 2.69 pixels in the vertical direction. In addition, the object detection speed also met the real-time requirements. The above results show that the proposed method has the potential to completely replace the manual interpretation and has very high application value.Gen ZhengJianhu ZhaoShaobo LiJie FengMDPI AGarticlesub-bottom profilerpipeline detectionYOLOv5szero-shotScienceQENRemote Sensing, Vol 13, Iss 4401, p 4401 (2021)
institution DOAJ
collection DOAJ
language EN
topic sub-bottom profiler
pipeline detection
YOLOv5s
zero-shot
Science
Q
spellingShingle sub-bottom profiler
pipeline detection
YOLOv5s
zero-shot
Science
Q
Gen Zheng
Jianhu Zhao
Shaobo Li
Jie Feng
Zero-Shot Pipeline Detection for Sub-Bottom Profiler Data Based on Imaging Principles
description With the increasing number of underwater pipeline investigation activities, the research on automatic pipeline detection is of great significance. At this stage, object detection algorithms based on Deep Learning (DL) are widely used due to their abilities to deal with various complex scenarios. However, DL algorithms require massive representative samples, which are difficult to obtain for pipeline detection with sub-bottom profiler (SBP) data. In this paper, a zero-shot pipeline detection method is proposed. First, an efficient sample synthesis method based on SBP imaging principles is proposed to generate samples. Then, the generated samples are used to train the YOLOv5s network and a pipeline detection strategy is developed to meet the real-time requirements. Finally, the trained model is tested with the measured data. In the experiment, the trained model achieved a mAP@0.5 of 0.962, and the mean deviation of the predicted pipeline position is 0.23 pixels with a standard deviation of 1.94 pixels in the horizontal direction and 0.34 pixels with a standard deviation of 2.69 pixels in the vertical direction. In addition, the object detection speed also met the real-time requirements. The above results show that the proposed method has the potential to completely replace the manual interpretation and has very high application value.
format article
author Gen Zheng
Jianhu Zhao
Shaobo Li
Jie Feng
author_facet Gen Zheng
Jianhu Zhao
Shaobo Li
Jie Feng
author_sort Gen Zheng
title Zero-Shot Pipeline Detection for Sub-Bottom Profiler Data Based on Imaging Principles
title_short Zero-Shot Pipeline Detection for Sub-Bottom Profiler Data Based on Imaging Principles
title_full Zero-Shot Pipeline Detection for Sub-Bottom Profiler Data Based on Imaging Principles
title_fullStr Zero-Shot Pipeline Detection for Sub-Bottom Profiler Data Based on Imaging Principles
title_full_unstemmed Zero-Shot Pipeline Detection for Sub-Bottom Profiler Data Based on Imaging Principles
title_sort zero-shot pipeline detection for sub-bottom profiler data based on imaging principles
publisher MDPI AG
publishDate 2021
url https://doaj.org/article/a9e6144e15a04f389958b3deb25ad5ce
work_keys_str_mv AT genzheng zeroshotpipelinedetectionforsubbottomprofilerdatabasedonimagingprinciples
AT jianhuzhao zeroshotpipelinedetectionforsubbottomprofilerdatabasedonimagingprinciples
AT shaoboli zeroshotpipelinedetectionforsubbottomprofilerdatabasedonimagingprinciples
AT jiefeng zeroshotpipelinedetectionforsubbottomprofilerdatabasedonimagingprinciples
_version_ 1718431664911155200