Underwater hyperspectral imaging for environmental mapping and monitoring of seabed habitats

Tegdan, Jorgen; Ekehaug, S.; Hansen, Ingrid Myrnes; Sandvik, Lars Martin; Steen, Kay Johnny; Pettersen, Ragnhild; Beuchel, Frank; Camus, Lionel

doi:10.1109/oceans-genova.2015.7271703

Cited by 25 publications

(20 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Variations in vehicle attitude may lead to distortions in the image, especially in across-track direction, resulting in a lower spatial resolution that can also influence spectral classification [62]. Without such variations, the overall data quality and spatial resolution of the UHI data in this study are higher than for previous UHI data sets [24]- [26].…”

Section: A Advantages and Disadvantages Of A Stationary Uhi Platformmentioning

confidence: 85%

“…The applied method was able to correct for most of the illumination influences, but residual illumination effects, and possibly influences from water column properties, may remain. The processed data therefore represent pseudo-reflectance rather than true reflectance data [24]. Although it would generally be preferable to have true reflectance data, this is not necessary for spectral classification.…”

Section: B Data Processingmentioning

confidence: 99%

“…The changing altitude was a challenge to data processing as it resulted in a systematic change in distance to the seafloor of both the UHI and the lamps, thus changing the illumination field along the track. In previous studies using ROVs [24]- [26], the altitude was generally kept constant to avoid changes in illumination conditions along the track. Correcting for the altitude changes therefore required a new processing approach, which was complicated by the absence of altitude measurements, swinging angle measurements, and an illumination reference, all of which had to be approximated.…”

Section: A Advantages and Disadvantages Of A Stationary Uhi Platformmentioning

confidence: 99%

“…Previous UHI surveys mainly focused on biological mapping, e.g., at cold-water coral reefs and kelp forests [24], [25]. However, the UHI method is also suitable for mapping and characterizing geological seafloor materials [23] and was recently applied to image manganese nodules in 4200 m water depth [26].…”

mentioning

confidence: 99%

See 3 more Smart Citations

Underwater Hyperspectral Imaging Using a Stationary Platform in the Trans-Atlantic Geotraverse Hydrothermal Field

Dumke

Ludvigsen

Ellefmo

et al. 2019

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

Underwater hyperspectral imaging is a relatively new method for characterizing seafloor composition. To date, it has been deployed from moving underwater vehicles, such as remotely operated vehicles and autonomous underwater vehicles. While moving vehicles allow relatively rapid surveying of several 10-1000 m 2 , they are subjected to short-term variations in vehicle attitude that often compromise image acquisition and quality. In this study, we tested a stationary platform that was landed on the seabed and used an underwater hyperspectral imager (UHI) on a vertical swinging bracket. The imaged seafloor areas have dimensions of 2.3 m × 1 m and are characterized by very stable UHI data of high spatial resolution. The study area was the TransAtlantic Geotraverse hydrothermal field at the Mid-Atlantic Ridge (26°N) in water depths of 3530-3660 m. UHI data were acquired a 12 stations on an active and an inactive hydrothermal sulfide mound. Based on supervised classification, 24 spectrally different seafloor materials were detected, including hydrothermal and non-hydrothermal materials, and benthic fauna. The results show that the UHI data are able to spectrally distinguish different types of surface materials and benthic fauna in hydrothermal areas, and may therefore represent a promising tool for high-resolution seafloor exploration in potential future deep-sea mining areas.

show abstract

Section: A Advantages and Disadvantages Of A Stationary Uhi Platformmentioning

confidence: 85%

Section: B Data Processingmentioning

confidence: 99%

Section: A Advantages and Disadvantages Of A Stationary Uhi Platformmentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

Underwater Hyperspectral Imaging Using a Stationary Platform in the Trans-Atlantic Geotraverse Hydrothermal Field

Dumke

Ludvigsen

Ellefmo

et al. 2019

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Thus, ground materials can be more discriminative with detailed spectral information (Ghamisi, Plaza, Chen, Li, & Plaza, 2017). Hyperspectral image classification has been developed for a variety of applications (Bioucas-Dias et al, 2013), such as environmental monitoring (Tegdan et al, 2015) and precision agriculture (Lacar, Lewis, & Grierson, 2001). However, due to the complexity of spectral and spatial structures, high dimensionality and strong correlation between adjacent bands, the classification of the hyperspectral image still remains a challenging task (Gomez-Chova, Tuia, Moser, & Camps-Valls, 2015).…”

Section: Introductionmentioning

confidence: 99%

Convolutional neural networks and local binary patterns for hyperspectral image classification

Wei

Liu

et al. 2019

European Journal of Remote Sensing

View full text Add to dashboard Cite

Convolutional neural networks (CNNs) have strong feature extraction capability, which have been used to extract features from the hyperspectral image. Local binary pattern (LBP) is a simple but powerful descriptor for spatial features, which can lessen the workload of CNNs and improve the classification accuracy. In order to make full use of the feature extraction capability of CNNs and the discrimination of LBP features, a novel classification method combining dual-channel CNNs and LBP is proposed. Specifically, a one-dimensional CNN (1D-CNN) is adopted to process original hyperspectral data to extract hierarchical spectral features and another same 1D-CNN is applied to process LBP features to further extract spatial features. Then, the concatenation of two fully connected layers from the two CNNs, which fused features, is fed into a softmax classifier to complete the classification. The experimental results demonstrate that the proposed method can provide 98.52%, 99.54% and 99.54% classification accuracy on the Indian Pines, University of Pavia and Salinas data, respectively. And the proposed method can also obtain good performance even with limited training samples.

show abstract

Parameter feature extraction for hyperspectral detection of the shallow underwater target

Xiong

2021

Sci. China Technol. Sci.

View full text Add to dashboard Cite

Underwater hyperspectral imaging for environmental mapping and monitoring of seabed habitats

Cited by 25 publications

References 3 publications

Underwater Hyperspectral Imaging Using a Stationary Platform in the Trans-Atlantic Geotraverse Hydrothermal Field

Underwater Hyperspectral Imaging Using a Stationary Platform in the Trans-Atlantic Geotraverse Hydrothermal Field

Convolutional neural networks and local binary patterns for hyperspectral image classification

Parameter feature extraction for hyperspectral detection of the shallow underwater target

Contact Info

Product

Resources

About