GEOSCAN, résultats de la recherche


TitreAdvanced approaches for video footage and photograph analysis in the Bay of Fundy
AuteurJerosch, K; Kostylev, V E; Todd, B A
Source36th Annual Colloquium & Annual General Meeting, abstracts; 2010.
LiensOnline - En ligne
Séries alt.Secteur des sciences de la Terre, Contribution externe 20100456
Réunion36th Annual Colloquium & Annual General Meeting; Wolfville; CA; février 6, 2010
SNRC21A; 21B; 21G; 21H
Lat/Long OENS -68.0000 -64.0000 46.0000 44.0000
Sujetsphotographie; photogéologie; taxonomie; géologie marine
ProgrammeGéoscience en mer, Énergie renouvelable offshore
Résumé(disponible en anglais seulement)
In 2009 the Bedford Institute for Oceanography (Natural Resources of Canada) generated a large quantity of footage - more than 25 km - in the Bay of Fundy on the 39th expedition of the "CCGS Hudson". Ecological marine research in epibenthic sea-floor communities relies strongly on the use of high resolution cameras analyzing the species composition and their correlation to depth and sediments, and their spatial distribution. The video material can currently, however, often not be used to its full potential as the analysis is time- and labor-intensive and requires the input of taxonomic experts. This poster presents two different approaches to make video data and still photographs of the Bay of Fundy available for an advanced analysis: GIDAS (Geospatial Image Database and Analysis System) is an automatic content-based analyses system for geo-referenced images. The automatic area recognition is achieved by learning visual properties of the different image regions. Therefore, each image is partitioned into a regular grid. Features are extracted and the classification is performed on the basis of these grid cells. Because of its ability working with geo-referenced images the analysis results are area accurate and can be exported as a shape-file. Biigle (Bielefeld Image Graphical Labeller and Explorer) is a Web 2.0 based platform containing easily uploaded images that can be accessed by collaborating scientists. The system provides not only basic image database exploration features but allows graphical annotation of semantic labels to image regions. Biigle also offers an application interface for machine-vision components aiming at the automated analysis of seafloor images. Laser point detection allows for an automated calibration of the area covered by the camera which is vital to derive faunal density estimates. This poster gives a report about strength and weakness of each system.