Multi-angle backscatter classification and sub-bottom profiling for improved seafloor characterization

Alevizos, Evangelos; Snellen, Mirjam; Simons, Dick G.; Siemes, Kerstin; Greinert, Jens

doi:10.1007/s11001-017-9325-4

Cited by 20 publications

(18 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The sediment samples of Table 1 were collected using a multi-corer and a Van-Veen grab. The selection of the training set was based on grain size analysis results and information from an unsupervised classification [3]. The unsupervised method applied in [3] yielded an optimum number of five classes therefore it was considered as guiding information for the supervised training in this study.…”

Section: Training Set Selectionmentioning

confidence: 99%

“…The selection of the training set was based on grain size analysis results and information from an unsupervised classification [3]. The unsupervised method applied in [3] yielded an optimum number of five classes therefore it was considered as guiding information for the supervised training in this study. The small amount of training set was chosen to fit the purpose of the study in assessing the ability of machine learning classifiers to perform adequate classification using a restricted amount of training set.…”

Section: Training Set Selectionmentioning

confidence: 99%

“…Seafloor acoustic mapping with multibeam echosounders (MBES) faces the emergence of new data acquisition styles, which also trigger the development of new data processing and interpretation approaches. Acquiring MBES datasets using multiple frequencies [1] or acquiring backscatter-dedicated MBES surveys [2,3] enormously increased the volume of data per seafloor area, leading to increased dimensionality of MBES datasets. At the same time, novel data processing methods appeared for improving the accuracy and detail of geo-information about the seafloor.…”

Section: Introductionmentioning

confidence: 99%

“…In addition, the separability of backscatter responses has an impact on discrimination of various seafloor types. Another problem considered in the studies of References [3,6] is the performance of traditional supervised classifiers using a restricted amount of ground truth data along with high-dimensional spatial datasets. The first two issues are highlighted by References [5,10,11].…”

Section: Introductionmentioning

confidence: 99%

“…Additionally, the resulting classification maps can be constructed with a spatial resolution comparable to that of ground truth data and the MBES footprint (meter scale), a fact that increases their validity and usefulness for the end-users [12]. Since the implementation of Reference [11], the above concept of an n-dimensional backscatter matrix was only later employed again, in the studies of References [3,13], where they took advantage of the high density of MBES soundings resulting from a backscatter-dedicated survey with >100% overlap. Interpolation of backscatter values from individual incidence angles was used for producing angular backscatter layers which were then stacked together to form the so-called hyper-angular cube matrix (HAC), which is presented here (Figure 1).…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

The Hyper-Angular Cube Concept for Improving the Spatial and Acoustic Resolution of MBES Backscatter Angular Response Analysis

Alevizos

Greinert

2018

Geosciences

Self Cite

View full text Add to dashboard Cite

This study presents a novel approach, based on high-dimensionality hydro-acoustic data, for improving the performance of angular response analysis (ARA) on multibeam backscatter data in terms of acoustic class separation and spatial resolution. This approach is based on the hyper-angular cube (HAC) data structure which offers the possibility to extract one angular response from each cell of the cube. The HAC consists of a finite number of backscatter layers, each representing backscatter values corresponding to single-incidence angle ensonifications. The construction of the HAC layers can be achieved either by interpolating dense soundings from highly overlapping multibeam echo-sounder (MBES) surveys (interpolated HAC, iHAC) or by producing several backscatter mosaics, each being normalized at a different incidence angle (synthetic HAC, sHAC). The latter approach can be applied to multibeam data with standard overlap, thus minimizing the cost for data acquisition. The sHAC is as efficient as the iHAC produced by actual soundings, providing distinct angular responses for each seafloor type. The HAC data structure increases acoustic class separability between different acoustic features. Moreover, the results of angular response analysis are applied on a fine spatial scale (cell dimensions) offering more detailed acoustic maps of the seafloor. Considering that angular information is expressed through high-dimensional backscatter layers, we further applied three machine learning algorithms (random forest, support vector machine, and artificial neural network) and one pattern recognition method (sum of absolute differences) for supervised classification of the HAC, using a limited amount of ground truth data (one sample per seafloor type). Results from supervised classification were compared with results from an unsupervised method for inter-comparison of the supervised algorithms. It was found that all algorithms (regarding both the iHAC and the sHAC) produced very similar results with good agreement (>0.5 kappa) with the unsupervised classification. Only the artificial neural network required the total amount of ground truth data for producing comparable results with the remaining algorithms.

show abstract

Section: Training Set Selectionmentioning

confidence: 99%

Section: Training Set Selectionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

The Hyper-Angular Cube Concept for Improving the Spatial and Acoustic Resolution of MBES Backscatter Angular Response Analysis

Alevizos

Greinert

2018

Geosciences

Self Cite

View full text Add to dashboard Cite

show abstract

A framework to quantify uncertainties of seafloor backscatter from swath mapping echosounders

2018

View full text Add to dashboard Cite

Multibeam echosounders (MBES) have become a widely used acoustic remote sensing tool to map and study the seafloor, providing co-located bathymetry and seafloor backscatter. Although the uncertainty associated with MBES-derived bathymetric data has been studied extensively, the question of backscatter uncertainty has been addressed only minimally and hinders the quantitative use of MBES seafloor backscatter. This paper explores approaches to identifying uncertainty sources associated with MBES-derived backscatter measurements. The major sources of uncertainty are catalogued and the magnitudes of their relative contributions to the backscatter uncertainty budget are evaluated. These major uncertainty sources include seafloor insonified area (1-3 dB), absorption coefficient (up to > 6 dB), random fluctuations in echo level (5.5 dB for a Rayleigh distribution), and sonar calibration (device dependent). The magnitudes of these uncertainty sources vary based on how these effects are compensated for during data acquisition and processing. Various cases (no compensation, partial compensation and full compensation) for seafloor insonified area, transmission losses and random fluctuations were modeled to estimate their uncertainties in different scenarios. Uncertainty related to the seafloor insonified area can be reduced significantly by accounting for seafloor slope during backscatter processing while transmission losses can be constrained by collecting full water column absorption coefficient profiles (temperature and salinity profiles). To reduce random fluctuations to below 1 dB, at least 20 samples are recommended to be used while computing mean values. The estimation of uncertainty in backscatter measurements is constrained by the fact that not all instrumental components are characterized and documented sufficiently for commercially available MBES. Further involvement from manufacturers in providing this essential information is critically required.

show abstract

Introduction to the Special Issue “Seafloor backscatter data from swath mapping echosounders: from technological development to novel applications”

Lamarche

Lurton

2018

Mar Geophys Res

View full text Add to dashboard Cite

Multi-angle backscatter classification and sub-bottom profiling for improved seafloor characterization

Cited by 20 publications

References 38 publications

The Hyper-Angular Cube Concept for Improving the Spatial and Acoustic Resolution of MBES Backscatter Angular Response Analysis

The Hyper-Angular Cube Concept for Improving the Spatial and Acoustic Resolution of MBES Backscatter Angular Response Analysis

A framework to quantify uncertainties of seafloor backscatter from swath mapping echosounders

Introduction to the Special Issue “Seafloor backscatter data from swath mapping echosounders: from technological development to novel applications”

Contact Info

Product

Resources

About