Elias Fakiris

Exploiting Environmental Information for Improved Underwater Target Classification in Sonar Imagery

In many remote-sensing applications, measured data are a strong function of the environment in which they are collected. This paper introduces a new context-dependent classification algorithm to address and exploit this phenomenon. Within the proposed framework, an ensemble of classifiers is constructed, each associated with a particular environment. The key to the method is that the relative importance of each object (i.e., data point) during the learning phase for a given classifier is controlled via a modulating factor based on the similarity of auxiliary environment features. Importantly, the number of classifiers to learn and all other associated model parameters are inferred automatically from the training data. The promise of the proposed method is demonstrated on classification tasks seeking to distinguish underwater targets from clutter in synthetic aperture sonar imagery. The measured data were collected with an autonomous underwater vehicle during several large experiments, conducted at sea between 2008 and 2012, in different geographical locations with diverse environmental conditions. For these data, the environment was quantified by features (extracted from the imagery directly) measuring the anisotropy and the complexity of the seabed. Experimental results suggest that the classification performance of the proposed approach compares favorably to conventional classification algorithms as well as state-of-the-art context-dependent methods. Results also reveal the object features that are salient for performing target classification in different underwater environments.

Illegal trawling and induced invasive algal spread as collaborative factors in a Posidonia oceanica meadow degradation

Posidonia oceanica, a key seagrass species of the Mediterranean Sea, shows clear signs of regression throughout the Mediterranean and illegal trawling is recognized as one of the main causes. We examined the condition of a P. oceanica meadow in Alykes Bay (Zakynthos Island, western Greece), a typical Mediterranean littoral area where illegal trawling is common practice, in respect to the total area affected, and in terms of possible ecological substitution. A side scan sonar (SSS) survey of the seafloor provided an image of the condition of the meadow and biological sampling evaluated the ecological status in affected meadow areas. SSS images revealed that trawling has a serious effect on the meadow, with 11% of the vegetated area being abraded, and the affected areas were also found to be fully colonized by the invasive alga Caulerpa racemosa. Moreover, unusually high densities of the polychaete Sabella pavonina were detected in the affected areas among C. racemosa fronds. Recolonization by P. oceanica of the affected meadow areas that have been colonized by C. racemosa seems improbable considering the allelopathic interactions between the species, with the alga displaying phytotoxic properties through caulerpenyne production and deterioration of the sediment quality.

Bed diversity in the shallow water environment of Pappas lagoon in Greece

Acoustic classification systems and the Sediment Trend Analysis method were used to identify and map the bed diversity in a very shallow (<3.0m), coastal lagoon, Pappas lagoon, Western Greece. Analogue acoustic data, collected by means of a 100 kHz side scan sonar system, were digitized and classified into six acoustic classes using recently developed acoustic classification systems (SonarClass, TargAn). By comparing the acoustic classes to ground truth data consisting of sediment grain size and visual inspection of the lagoon-bed, it is demonstrated that the six acoustic classes correlate well with the predominant surface sediment types and vegetation. Thus the spatial distribution of the classes can be considered to represent the spatial pattern of the sedimentary assemblages of the lagoon. The grain size trend analysis identified three dominant sediment pathways and directional trends which could be related to the predominant wind direction, the sediment influx through the inlets and the sediment supply from a small stream in the southern part of the lagoon. The integration of acoustic and sedimentological data together with advanced data processing systems leads to a better understanding of the sedimentary, morphological and biological processes in a shallow lagoon in different spatial and temporal scales and will therefore be beneficial to both sedimentological and biotic-diversity studies.

An Automatic Target Detection Algorithm for Swath Sonar Backscatter Imagery, Using Image Texture and Independent Component Analysis

In the present paper, a methodological scheme, bringing together common Acoustic Seabed Classification (ASC) systems and a powerful data decomposition approach, called Independent Component Analysis (ICA), is demonstrated regarding its suitability for detecting small targets in Side Scan Sonar imagery. Traditional ASC systems extract numerous texture descriptors, leading to a large feature vector, the dimensionality of which is reduced by means of data decomposition techniques, usually Principal Component Analysis (PCA), prior to classification. However, in the target detection issue, data decomposition should point towards finding components that represent sub-ordinary image information (i.e., small targets) rather than a dominant one. ICA has long been proved to be suitable for separating targets from a background, and this study represents a novel exhibition of its applicability to Side Scan Sonar (SSS) images. The present study attempts to build a fully automated target detection approach that combines image based feature extraction, ICA, and unsupervised classification. The suitability of the proposed approach has been demonstrated using an SSS data-set containing more than 70 manmade targets, most of them metallic, validated through a marine magnetic survey or ground truthing inspection. The method exhibited very good performance as it was able to detect more than 77% of the targets and it produced less than seven false alarms per km2. Moreover, it was compared to cases where, in the exact same methodological scheme, no decomposition technique is used, or PCA is employed instead of ICA, achieving the highest detection rate, but, more importantly, producing more than six times less false alarms, thus proving that ICA successfully manages to maximize target to background separation.

Litter composition and source contribution for 80 beaches in Greece, Eastern Mediterranean: A nationwide voluntary clean-up campaign

Litter that accumulates in the marine environment is recognized as a global environmental problem with numerous adverse effects. Despite Greece having more than 16,000 km of coastline, studies on litter composition and sources on its beaches are scarce and fragmentary. Therefore, this study aimed to quantify beach-stranded litter by type of material (plastic, glass, paper, aluminum, other metals, building materials, rope) on 80 Greek beaches, and to explore the relationships between different litter materials and marine- and land-based sources. Field work and data collection were accomplished with the participation of 10,938 and 15,748 volunteers in 2006 and 2007, respectively, and were integrated in the nationwide campaign of voluntary beach clean-ups: “Clean up the Med.” The participation of thousands of volunteers of different ages and social backgrounds in the data collection process not only led to the realization of a national level survey, but also played a significant role in raising public awareness on marine environmental issues, and in promoting a sense of responsibility in protecting marine resources. Plastic was the most abundant litter material found on the surveyed beaches (43–51%), followed by paper (13–18%) and aluminum (7–12%). Onshore and nearshore recreational activities, such as bathing and recreational boating, appeared to be the most dominant litter sources, as they affect the majority (56) of the surveyed beaches, suggesting that environmental education and awareness could reduce the beach-litter pollution problem. Navigation seemed to be the dominant source of marine-based litter, affecting more than half (46) of the surveyed beaches.

Uranium and other natural radionuclides in the sediments of a Mediterranean fjord-like embayment, Amvrakikos Gulf (Ionian Sea), Greece.

The distribution of the natural radionuclides ((238)U, (232)Th, (226)Ra, (40)K) and the artificial (137)Cs was studied in sediment cores collected from Amvrakikos Gulf, a seasonal anoxic marine basin, using γ-ray spectrometry. The activity of radionuclides, along with the concentrations of Fe and Mn, were also studied in relation to the total organic carbon and the granulometric fractions of the sediments. The results obtained revealed higher (238)U activity concentrations in all the examined sediment samples compared to the world and Greek average values for soil. The high activity values of (238)U are attributed, besides the lattice-held fraction, to phosphate fertilizer inputs in the Gulf via major rivers and/or to alteration processes of phosphate ores located mainly in the drainage basin of the river Louros. The elevated activity values of (40)K could be attributed to the mineralogical composition of the sediments and to phosphate fertilizers containing potassium. Organic matter seems to be a more efficient sorbent for U than clay minerals and amorphous Fe and Mn-oxyhydroxides. Scanning electron microscopy, together with qualitative analysis of some smectites, reveals the occurrence of U, suggesting a limited absorption of U onto clay minerals. The applied BCR sequential extraction procedure revealed that U was found mainly in the refractory phase or associated with organic matter and to a lesser extent as surface-coating oxides, with the exception of one sediment core which is characterized by high content of fresh marine organic matter and presents high percentage of U in the exchangeable fraction.

Beach litter dynamics on Mediterranean coasts: Distinguishing sources and pathways

We assessed amounts, composition and net accumulation rates every ~15days of beach macro litter (≥2.5cm) on 4 Mediterranean beaches, on Corfu island, N. Ionian Sea, taking into account natural and anthropogenic drivers. Average net accumulation rate on all beaches was found 142±115N/100m/15d. By applying a Generalized Linear Model (GzLM) it was shown that sea transport is the dominant pathway affecting the amount and variability in beach litter loadings. Principal Component Analysis (PCA) on compositional data and indicator items discerned two more pathways of beach litter, i.e. in situ litter from beach goers and wind and/or runoff transport of litter from land. By comparing the PCA results to those from a simple item to source attribution, it is shown that regardless their source litter items arrive at beaches from various pathways. Our data provide baseline knowledge for designing monitoring strategies and for setting management targets.

Quantification of regions of interest in swath sonar backscatter images using grey-level and shape geometry descriptors: the TargAn software

In this paper, the TargAn software package that is dedicated to parameterizing regions of interest (ROIs) in greyscale images that reflect backscatter information derived by marine geo-acoustical instrumentation (e.g. Side Scan Sonar and Multi-Beam Echo-Sounder) is presented. The ROIs, whose boundaries are marked out either manually or via simple segmentation techniques, are analyzed for as many as 37 features. The adopted and developed methodologies lead to the extraction of: (1) grey-level intensity (1st order) and texture analysis statistics estimated from the inner ROI, (2) descriptors that measure the separation of the ROI in comparison to the intensity characteristics of the peripheral seabed, (3) shape geometry descriptors of the ROI’s boundary itself and (4) regional statistics of distinct (segmented) objects possibly included in the ROI. TargAn is implemented in Matlab with a graphical user interface that helps the user to have control over the digitization, segmentation and feature extraction processes involved. It also provides tools for the construction of compact geo-databases, suitable for geostatistical analysis and visualization in popular Geographical Information Systems, concerning the extracted descriptors and the geographic features (e.g. ROIs’ boundaries, skeletons, segmented objects) that have been considered for ROIs’ analyses. The TargAn software is particularly useful when large amounts of image ROIs need to be objectively quantified and is demonstrated through two case studies regarding Side Scan Sonar imageries. The first one concerns the quantification of marine biohabitats (coralline formations) while the second exhibits the geometrical analysis of pockmarks.

Integrating dual frequency side-scan sonar data and multibeam backscatter, angular response and bathymetry, for benthic habitat mapping in the Laganas Gulf MPA, Zakinthos Isl., Greece

The preferred procedure nowadays for benthic habitat mapping is combining marine acoustic and ground truthing methods, with the former ones tending to be the swath sonars, such as the Multi Beam Echo Sounders (MBES) and the Side-Scan Sonars (SSS). Both above acoustic systems, in conjunction with an extensive underwater video footage, were employed to map in detail benthic habitats in the marine area of Laganas Gulf in the National Marine Park of Zakynthos Isl., Greece, including key protected habitats such as P.oceanica beds and coralligenous formations. Object oriented seafloor classification was achieved taking advantage of the multi-layer information available, including the two individual frequencies of SSS and the MBES backscatter, angular response, and bathymetry data. The extracted statistical derivatives regarded: (1) textural analysis of the multi-frequency backscatter imagery, (2) angular range analysis of the MBES backscatter, as well as (3) application of various bathymetric indices. Those der...

The Disappearance of Helike-Classical Greece—New Remote Sensing and Geological Evidence

Helike, the Achaean Dodecapolis capital, in the Corinth Gulf, Greece, was, according to historical sources, devastated and disappeared from sight during an earthquake followed by sea invasion on to land in 373/372 B.C. A marine remote sensing survey, which was carried out to examine a landslide following a 6.2 R earthquake in 1995, that affected the coastal and near-shore delta plain zone on which Helike stood, accompanied by onshore borehole data, enabled us to postulate the geological processes leading to the Helike catastrophe. Helike was initially leveled during a 6 to 6.7 R earthquake; it is postulated to have then submerged following a translational landslide caused by liquefaction. This Helike catastrophe model is consistent with historical sources and current views regarding the expected geological hazards magnitude in the Gulf.