Geospatial intelligence and visual classification of environmentally observed species in the future Internet

Abstract

The rapid development of advanced smart communication tools with good quality and resolution video cameras, audio and GPS devices in the last few years shall lead to profound impacts on the way future environmental
observations are conducted and accessed by communities. The resulting large scale interconnections of these "Future Internet Things" form a large environmental sensing network which will generate large volumes of quality
environmental observations and at highly localised spatial scales. This enablement in environmental sensing at local scales will be of great importance to contribute in the study of fauna and flora in the near future, particularly
on the effect of climate change on biodiversity in various regions of Europe and beyond. The Future Internet could also potentially become the de facto information space to provide participative real-time sensing by communities
and improve our situation awareness of the effect of climate on local environments. In the ENVIROFI(2011-2013) Usage Area project in the FP7 FI-PPP programme, a set of requirements for specific (and generic) enablers is
achieved with the potential establishment of participating community observatories of the future. In particular, the specific enablement of interest concerns the building of future interoperable services for the management of
environmental data intelligently with tagged contextual geo-spatial information generated by multiple operators in communities (Using smart phones). The classification of observed species in the resulting images is achieved
with structured data pre-processing, semantic enrichment using contextual geospatial information, and high level fusion with controlled uncertainty estimations. The returned identification of species is further improved using
future ground truth corrections and learning by the specific enablers.

More information

Identifiers

ORCID for Z. Sabeur: orcid.org/0000-0003-4325-4871

Catalogue record

Export record