PostDoc Job: GPU Computational Scaling to Support Emergency Mapping
The current vacancy is in the Institute for the Protection and Security of the Citizen (located in Ispra, Italy). The Institutes provides research results and supports EU policy-makers in their effort towards global security and protection of European citizens from accidents, deliberate attacks, fraud and illegal actions against EU policies. In relation to the workplace general information can be found at the following website: http://ec.europa.eu/dgs/jrc/index.cfm?id=5790 .The vacancy is within the Global Security and Crisis Management unit.
The Global Security and Crisis Management unit is assisting the Commission’s Directorate General Enterprise and Industry in the implementation of the GMES initial operations Emergency Management Service (GIO-EMS). This service is aimed at providing geospatial information derived from GMES space assets, in particular Earth Observation satellites, to support decision making in the various phases of the emergency management cycle. Whereas GIO-EMS has a strict operational character, a number of research topics are aligned with potential enhancements to deployed methodologies, integration of ancillary geospatial data resources, sampling issues and uptake and interaction with evolving “crowd-sourced” mapping efforts.
The Unit’s ISFEREA action is focused on the integration of advanced image processing methodologies into GIO-EMS-like workflows. Besides research developments in image morphology operators, segmentation and spectral and spatial characterization of image objects, the focus has been on scaling up computational performance to match near-real-time information needs in crisis and disaster contexts. In particular, GPU computing based implementations have demonstrated significant results. We are looking for a candidate to expand upon our GPU experience to scale up current in-house developed fast processing workflows. These may include, but are not limited to, fast image registration, change detection, object delineation, segmentation and classification. The work requires the integration and extension of GPU-accelerated image processing routines, multiple-GPU parallelisation, benchmarking and quality of service review and supporting the team in the operation of dedicated image processing workflows. Other areas of research will deal with the integration of geospatial data sources to steer and constrain operations and make use of object data bases for recognition tasks. The research will contribute to other areas, such as image interpretation, sampling and geo-statistical analysis.
The work is highly practical and goal-oriented. Research results are expected to be used operationally. Applicants should therefore have a strong programming background combined with knowledge of geospatial analysis and image processing. The candidate is also expected to contribute to publications of scientific research results.
The working language of the action is English, so strong English language communication and writing skills are required.
Tasks:
• Carry out research and development of GPU-accelerated image processing algorithms that are relevant in operational crisis and disaster mapping contexts;
• Scale relevant image processing tasks to multi-GPU High Performance Computing systems;
• Integrate geospatial information and object data bases to steer GPU-enabled workflow;
• Benchmark and compare GPU implementations to conventional approaches;
• Assist team members to make practical use of GPU-enabled workflow in on-demand real time applications;
• Contribute to team efforts in sampling issues, validation, high volume data processing, visualisation;
• Author and/or contribute to peer-reviewed scientific manuscripts and present results at relevant meetings and fora.
Required profile:
The system within which the results will be deployed is based on the Linux Operating System and NVIDIA Tesla GPU computing hardware. Therefore, the successful candidate should have or demonstrate:
• Excellent programming skills in C/C++ combined with CUDA and/or OpenCL;
• At least 2 years of relevant experience with GPU processing of large image data sets, preferably derived from remote sensing platforms;
• Excellent software design and software documentation skills;
• Very good knowledge of image processing algorithms, preferably those relevant to remote sensing analysis;
• Working knowledge of Linux and related scripting languages (bash, python, perl) is required;
• Working knowledge of (open source) Geographical Information Systems and (geo)spatial data base and analysis software is an asset;
• Working knowledge of Java and/or Groovy is an asset;
• A scientific publication record, preferably including submissions to peer reviewed journals, is an asset.
• Ability to work in a multi-disciplinary and multi-cultural team
• Good analytical and problem solving skills
• Very good spoken/written communication and good interpersonal skills.
Requirements: The successful candidate will have a university degree in computer science or signal processing, plus a doctoral degree (PhD) in a similar discipline or, if not having the PhD, a minimum 5 years of full time related research experience after the first University degree giving access to doctoral (PhD) studies.
Other requirements: according to the eligibility requirements for the grantholder category (please see: http://ipsc.jrc.ec.europa.eu/fileadmin/repository/jobs/AdminrulesGholders_2010_Ares_2010_85578.pdf)
In order to apply, the following documents should be sent to JRC-IPSC-GRANTHOLDERS@ec.europa.eu :
-
updated CV in English
-
application form (3 pages) please sign the application form by hand
-
copy of your ID card/passport
Duration: 36 months
Location: Ispra, Italy
Category: Jobs
Subscribe
If you enjoyed this article, subscribe to receive more just like it.
