Towards Underwater Robust Image Transmission Using Acoustic Communications

被引：0

作者：

Gonzalez, Caterina Muntaner ^{[1
]}

Nordfeldt-Fiol, Bo Miquel ^{[1
]}

Abadal, Miguel Martin ^{[1
]}

Torres, Antoni Martorell ^{[1
]}

Bonin-Font, Francisco ^{[1
]}

Arcas, Fernando Molina ^{[2
]}

Prades, Raul Marin ^{[2
]}

Oliver, Gabriel ^{[1
]}

Valero, Pedro J. Sanz ^{[2
]}

机构：

[1] Univ Balear Islands UIB, Syst Robot & Vision Grp, Palma de Mallorca, Spain

[2] Univ Jaume I UJI, Interact & Robot Syst Lab, Castellon de La Plana, Spain

来源：

OCEANS 2023 - LIMERICK | 2023年

关键词：

Compression; Underwater Autonomous Robots; Acoustic Underwater Data Transmission; Object Detection; COMPRESSION; REGION;

D O I：

10.1109/OCEANSLimerick52467.2023.10244660

中图分类号：

U6 [水路运输]; P75 [海洋工程];

学科分类号：

0814 ; 081505 ; 0824 ; 082401 ;

摘要：

This paper presents a novel approach aimed at broadcasting pre-compressed images of reference hot-spots between marine robots through an acoustic link. The transmission must be robust, even on areas with potential acoustic multi-scattering. Moreover, the system should maintain enough resolution to make the exchanged images useful for applications that imply or require automatic object detection, remote visualization, robust feature matching and reliable registrations. The acoustic exchange of images is framed in the context of a Spanish project in which online automatic remote awareness of underwater pipeline structural elements is required to identify and locate the site for a multi-robot cooperative intervention. Experiments in real marine scenarios include detection of underwater pipeline valves using deep learning on images, their compression and acoustic transmission between two autonomous marine robots with high success rates.

引用

页数：8