En poursuivant votre navigation sur ce site, vous acceptez le dépôt de cookies dans votre navigateur. (En savoir plus)
Portail > Offres > Offre UPR8001-ARIHER-002 - Post-doc H/F : Développement de fonctionnalités de vision embarquées sur aéronef, pour l'assistance à la navigation dans les zones aéroportuaires, par méthodes d'apprentissage

Development of embedded vision functionalities on aircraft, for navigation assistance in airport areas, using learning methods

This offer is available in the following languages:
Français - Anglais

Date Limite Candidature : mercredi 1 juin 2022

Assurez-vous que votre profil candidat soit correctement renseigné avant de postuler. Les informations de votre profil complètent celles associées à chaque candidature. Afin d’augmenter votre visibilité sur notre Portail Emploi et ainsi permettre aux recruteurs de consulter votre profil candidat, vous avez la possibilité de déposer votre CV dans notre CVThèque en un clic !

General information

Reference : UPR8001-ARIHER-002
Workplace : TOULOUSE
Date of publication : Wednesday, May 11, 2022
Type of Contract : FTC Scientist
Contract Period : 20 months
Expected date of employment : 1 September 2022
Proportion of work : Full time
Remuneration : between 2663 and 3069 euros gross per month depending on experience
Desired level of education : PhD
Experience required : Indifferent


The candidate will participate in the design of a perception system that will eventually be integrated on a heterogeneous architecture (multi-core CPU, and GPU...), based on an algorithmic specification to achieve good performance with strong real-time constraints.
He will participate in the study and development of image processing algorithms, algorithms that will aim to characterize and model the sensors to be embarked on an aircraft, and to provide monitoring functions to avoid any runway excursion during the taxiing phase on runways and taxiways. Therefore these algorithms will have two objectives:
1. to estimate the intrinsic capabilities of the sensors, e.g. maximum detection range of runway markings and edges, depending on the positioning of the sensor on the aircraft, visibility conditions, etc.
2. Measure the performance of prototyped functions for migration to an embedded architecture while respecting real-time constraints.

These functions will deal with the detection and tracking of markings, runway edges, etc. using geometric approaches or classical learning (classification of regions of interest based on descriptors) or deep learning (CNN). The algorithms will be trained and validated on the basis of mixed data sets composed of synthetic images automatically annotated by exploiting the position of the sensors in the scene and the airport model, and real images annotated approximately via classical treatments (color, gradient...). An important subject, treated in connection with the company which provides the generator of synthetic images, will concern the formation of the data set, the methods of data augmentation, the distribution synthetic images / real images, etc...

The candidate will work with the permanent staff of the RAP team of LAAS-CNRS involved in this project: he/she will be in charge of its technical management (documentation, deliverables, exchanges with partners, etc.). He/she will have to participate (1) in the project meetings, to present the results obtained by the team to the scientific experts of the DGAC, and (2) in the dissemination of the scientific results during scientific conferences and workshops.
He/she will have to cooperate with an engineer recruited on the same project who will take care of the integration part, and with people already present in the team, who have related subjects (obstacle detection by deep learning, detection/tracking of markings by particle filtering, certification of vision algorithms...); in this team, he/she will be more in charge of the development of algorithms and their evaluation against existing methods available in Open Source, on the basis of public data sets


-State of the art and contributions on 3D vision.
-Development of algorithms in C/C++/Python, possibly using Matlab prototyping.
-Technical management of the project (documentation, organization, tests, formal presentations...)


- PhD. in computer vision, learning, for applications in Robotics or "Intelligent" Vehicles.
- 3D vision (projective geometry, camera model, calibration...)
- Tools for vision and machine learning: OpenCV, MATLAB, CNN libraries and architectures...
- Autonomy, team work
- Project management
- Writing skills

Work Context

This job is part of a research project with a consortium coordinated by AIRBUS. This project, financed by the French Civil Aviation Authority (DGAC), aims to develop a system to assist the piloting of an aircraft in airport areas, assistance especially required in poor visibility conditions (night, fog, rain, etc.). Different multi-spectral images are acquired and analyzed, in order to :
- on the one hand, to generate a synthetic and realistic view of the environment in which an aircraft evolves. This view, augmented with symbolic and textual information, is displayed in the cockpit; the purpose is to inform the pilot of any potential risk (situational awareness) during the navigation phase on the taxiways from the parking area to the runway, and vice versa;
- on the other hand, to generate alerts to the aircraft's control system, which will eventually be equipped with sufficient autonomy to decide on the actions to be taken in these situations.

We talk about it on Twitter!