Simulator-aided Edge-based Acoustic Camera Pose Estimation

Wang, Yusheng; Ji, Yonghoon; Liu, Dingyu; Yamashita, Atsushi; Asama, Hajime

doi:10.1109/oceanschennai45887.2022.9775502

Cited by 6 publications

(2 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is also necessary to know the relative pose between the scene and the FLS. The initial value is measured by a diver using a ruler and the accurate result is obtained by comparing the synthetic image with the real image in the edge domain [29]. After acquiring one pose accurately, the rest can be calculated with the help of control input for the moving device and the rotator.…”

Section: B Real Experiments 1) Datasetmentioning

confidence: 99%

Learning Pseudo Front Depth for 2D Forward-Looking Sonar-based Multi-view Stereo

Wang

Asama

et al. 2022

2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

Self Cite

View full text Add to dashboard Cite

2D forward-looking sonar is a crucial sensor for underwater robotic perception. A well-known problem in this field is estimating missing information in the elevation direction during sonar imaging. There are demands to estimate 3D information per image for 3D mapping and robot navigation during fly-through missions. Recent learning-based methods have demonstrated their strengths, but there are still drawbacks. Supervised learning methods have achieved highquality results but may require further efforts to acquire 3D ground-truth labels. The existing self-supervised method requires pretraining using synthetic images with 3D supervision. This study aims to realize stable self-supervised learning of elevation angle estimation without pretraining using synthetic images. Failures during self-supervised learning may be caused by motion degeneracy problems. We first analyze the motion field of 2D forward-looking sonar, which is related to the main supervision signal. We utilize a modern learning framework and prove that if the training dataset is built with effective motions, the network can be trained in a self-supervised manner without the knowledge of synthetic data. Both simulation and real experiments validate the proposed method.

show abstract

Section: B Real Experiments 1) Datasetmentioning

confidence: 99%

Learning Pseudo Front Depth for 2D Forward-Looking Sonar-based Multi-view Stereo

Wang

Asama

et al. 2022

2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

Self Cite

View full text Add to dashboard Cite

show abstract

“…Acoustic landmarks were used for underwater navigation, and they were detected using a technique known as beam slice-based recognition [18]. An acoustic image simulator was used to determine the relative 3-D pose of a target whose shape and dimension are known, and it is subsequently employed for the global localization of underwater vehicles [19].…”

Section: Related Workmentioning

confidence: 99%

Robust Underwater Localization Using Acoustic Image Alignment for Autonomous Intervention Systems

Park

Kim

2022

IEEE Access

View full text Add to dashboard Cite

Acoustic imaging sonar can be used as important navigation sensors for underwater intervention systems as they consistently provide spatial information about the surrounding environment, even in limited visibility conditions. However, acoustic imaging sonars are known for high spatial ambiguity and low resolution of their measurements, which makes it challenging to obtain precise navigational information. This paper presents a novel localization method based on sensor fusion using an integrated IMU-DVL system and an acoustic imaging sonar. A sonar simulator is implemented and used to estimate the pose of the robot based on a single acoustic image, and it is formulated as an image alignment problem between a simulated acoustic image and an actual acoustic image. For this, an approximate nearest neighbor search method is employed for initial pose estimation, and a newly developed acoustic image alignment method is applied to obtain more accurate results in a continuous pose space. These methods are then combined to construct an integrated localization system using all the navigation sensors on the robot. Its feasibility and utility of the proposed approach is shown through an experimental validation in a test tank.

show abstract

Motion Degeneracy in Self-supervised Learning of Elevation Angle Estimation for 2D Forward-Looking Sonar

Wang,

Ji,

et al. 2023

2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

View full text Add to dashboard Cite

Simulator-aided Edge-based Acoustic Camera Pose Estimation

Cited by 6 publications

References 11 publications

Learning Pseudo Front Depth for 2D Forward-Looking Sonar-based Multi-view Stereo

Learning Pseudo Front Depth for 2D Forward-Looking Sonar-based Multi-view Stereo

Robust Underwater Localization Using Acoustic Image Alignment for Autonomous Intervention Systems

Motion Degeneracy in Self-supervised Learning of Elevation Angle Estimation for 2D Forward-Looking Sonar

Contact Info

Product

Resources

About