IMT Atlantique, internationally recognized for the quality of its research, is a leading general engineering school under the authority of the Ministry of Industry and Digital Technologies, ranked in the 3 main international rankings (THE, SHANGAI, QS). With three campuses in Brest, Nantes and Rennes, IMT Atlantique aims to combine digital technology and energy to transform society and industry through training, research and innovation.
The teaching and research Département Image et Traitement de l’Information is located on the Brest campus and covers several research areas. Among them is the knowledge of the marine environment, which is a major challenge of the 21st century, to meet the issues of the maritime industry economy, the coastal development and protection, along with fundamental aspects of security and safety of the territory.
The global observation of the maritime environment is based on the acquisition of data on a large time scale or in real time, by remote sensors. These data, representing a considerable volume, are visualized in the form of 2D/3D signals (spatial information) with a complementary component of temporal evolution (4D) and require the development of suitable processing methods to extract the relevant information.
Used sensors are active (such as radars, lidars or sonars) and their interactions with the environment modify the perception of it, or passive to collect additional information.
This position is open within the framework of the project Fusion and comparison of multi-sensor bathymetric data, financed by the Shom (hydrographic and oceanographic service of the Navy). The mission of the Shom is to know and describe the physical marine environment according to its relations with the atmosphere, the seabed and the coastal areas, to forecast its evolution and to ensure the dissemination of the corresponding information.
The Shom generates bathymetric products that are considered to be a very good representation of bathymetric knowledge adapted to particular applications and identified needs. Selection of the data that compose these products is currently done by means of expert rules, based essentially on characteristic metadata of a batch (unit of bathymetric information with the same hydrographic characteristics). Following this selection of data sources, a spatial interpolation process is applied to merge and provide an estimate of the bathymetry at any point in the model, for cartographic products or for digital terrain models.
The current approach does not include the multi-scale coherence (or lack thereof) of point clouds. Moreover, the increasing data volumes generated by new sensors imply a refinement of the methodology and tools in terms of both automatization and a more efficient decision support. We are looking for indicators/tools that allow a probe-based analysis, hybridizing the global qualitative approach (expert rules) with the local geostatistical approach.
The proposed work consists of several stages:
PhD obtained after February 2020 in Computer Science and Information Processing, or an engineering degree or equivalent in these two fields with two years of experience.