Evaluation of Deep Learning Strategies for Underwater Object Search

Knapik, Mateusz; Cyganek, Bogusław

doi:10.1109/sa47457.2019.8938092

Cited by 3 publications

(2 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This all leads to the eventual end goal of enabling constant, long-term monitoring with timely feedback to human operators through automated detection and classification of marine objects. Further, the integration of such a low-cost, high-performance sensor can form the basis for augmenting prior work on deep learning for recognition of marine objects [ 12 ].…”

Section: Introductionmentioning

confidence: 99%

Compact Quantum Magnetometer System on an Agile Underwater Glider

Page

Lambert

Mahmoudian

et al. 2021

Sensors

View full text Add to dashboard Cite

This paper presents results from the integration of a compact quantum magnetometer system and an agile underwater glider for magnetic survey. A highly maneuverable underwater glider, ROUGHIE, was customized to carry an increased payload and reduce the vehicle’s magnetic signature. A sensor suite composed of a vector and scalar magnetometer was mounted in an external boom at the rear of the vehicle. The combined system was deployed in a constrained pool environment to detect seeded magnetic targets and create a magnetic map of the test area. Presented is a systematic magnetic disturbance reduction process, test procedure for anomaly mapping, and results from constrained operation featuring underwater motion capture system for ground truth localization. Validation in the noisy and constrained pool environment creates a trajectory towards affordable littoral magnetic anomaly mapping infrastructure. Such a marine sensor technology will be capable of extended operation in challenging areas while providing high-resolution, timely magnetic data to operators for automated detection and classification of marine objects.

show abstract

Section: Introductionmentioning

confidence: 99%

Compact Quantum Magnetometer System on an Agile Underwater Glider

Page

Lambert

Mahmoudian

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…However, computational support is needed to make the best use of the vast amounts of data generated, e.g., by stationary underwater observatories [4] or seafloor observation systems [5]. A lot of work exists in the context of (semi-)automated detection and classification of species in marine images [3,[6][7][8][9][10][11][12]. All these works employ some kind of machine learning algorithm to render a data-driven model of the task to be performed (like object detection or classification).…”

Section: Introductionmentioning

confidence: 99%

ALMI—A Generic Active Learning System for Computational Object Classification in Marine Observation Images

Möller

Nattkemper

2021

Sensors

View full text Add to dashboard Cite

In recent years, an increasing number of cabled Fixed Underwater Observatories (FUOs) have been deployed, many of them equipped with digital cameras recording high-resolution digital image time series for a given period. The manual extraction of quantitative information from these data regarding resident species is necessary to link the image time series information to data from other sensors but requires computational support to overcome the bottleneck problem in manual analysis. As a priori knowledge about the objects of interest in the images is almost never available, computational methods are required that are not dependent on the posterior availability of a large training data set of annotated images. In this paper, we propose a new strategy for collecting and using training data for machine learning-based observatory image interpretation much more efficiently. The method combines the training efficiency of a special active learning procedure with the advantages of deep learning feature representations. The method is tested on two highly disparate data sets. In our experiments, we can show that the proposed method ALMI achieves on one data set a classification accuracy A > 90% with less than N = 258 data samples and A > 80% after N = 150 iterations, i.e., training samples, on the other data set outperforming the reference method regarding accuracy and training data required.

show abstract

An ensemble deep learning method with optimized weights for drone-based water rescue and surveillance

Ga̧sienica-Józkowy¹,

Knapik²,

Cyganek³

2021

ICA

View full text Add to dashboard Cite

Today’s deep learning architectures, if trained with proper dataset, can be used for object detection in marine search and rescue operations. In this paper a dataset for maritime search and rescue purposes is proposed. It contains aerial-drone videos with 40,000 hand-annotated persons and objects floating in the water, many of small size, which makes them difficult to detect. The second contribution is our proposed object detection method. It is an ensemble composed of a number of the deep convolutional neural networks, orchestrated by the fusion module with the nonlinearly optimized voting weights. The method achieves over 82% of average precision on the new aerial-drone floating objects dataset and outperforms each of the state-of-the-art deep neural networks, such as YOLOv3, -v4, Faster R-CNN, RetinaNet, and SSD300. The dataset is publicly available from the Internet.

show abstract

Evaluation of Deep Learning Strategies for Underwater Object Search

Cited by 3 publications

References 19 publications

Compact Quantum Magnetometer System on an Agile Underwater Glider

Compact Quantum Magnetometer System on an Agile Underwater Glider

ALMI—A Generic Active Learning System for Computational Object Classification in Marine Observation Images

An ensemble deep learning method with optimized weights for drone-based water rescue and surveillance

Contact Info

Product

Resources

About