Partenaires

CNRS
Ministère de l'Enseignement Supérieur et de la Recherche
Ecole Centrale de Lille
Université de Lille 1
CRIStAL

« mai 2017 »
L M M J V S D
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30 31 1 2 3 4


Rechercher

Sur ce site

Sur le Web du CNRS


Accueil du site > Soutenances de Thèses > DAWOOD Maya

Soutenance de Thèse Mlle DAWOOD Maya

Mlle DAWOOD Maya a soutenu sa thèse, le Vendredi 1er mars 2013 - Amphithéâtre 2A02 - IUT A de Lille - 9 h 30

Titre de la thèse :

"Vehicle geo-localization based on GPS, Vision and 3D virtual model"

Date :

Vendredi 1er mars 2013 - Amphithéâtre 2A02 - IUT A de Lille - 9 h 30

Le jury de thèse est composé de :

Directeur de thèse : Maan EL BADAOUI EL NAJJAR, Maître de Conférences - Université Lille 1
Directeur de thèse : Mohamad KHALIL, Professeur - Université Libanaise FG1
Directeur de thèse : Denis POMORSKI, Professeur - Université Lille 1
Rapporteur : Véronique BERGES-CHERFAOUI, Maître de Conférences - Université de Technologie de Compiègne
Rapporteur : Bassam DAYA, Professeur - Université Libanaise FG3
Membre : Abdelaziz BENSHRAIR, Professeur - INSA Rouen
Membre : Jamal CHARARA, Professeur - Université Libanaise FG2
Membre : Cindy CAPPELLE, Maître de Conférences - Université de Technologie de Belfort-Montbéliard
Membre : Bachar EL HASSAN, Maître de Conférences - Université Libanaise FG1

Résumé :

Vehicle geo-localization remains to be one of the challenging problems in urban areas. For this purpose, Global Positioning System (GPS) receiver is usually the main sensor. But, the use of GPS alone is not sufficient in many urban environments due to wave multi-path. In order to provide accurate and robust localization, GPS has to be helped with other sensors like dead-reckoned sensors, map data, cameras or LIDAR. In this thesis, a new observation of the absolute pose of the vehicle is proposed to back up GPS measurements. The proposed approach exploits virtual 3D city model managed by a 3D Geographical Information System (3D GIS) and a video camera. Vehicle geo-localization uses several sources of information : a GPS receiver, proprioceptive sensors (odometers and gyrometer), a video camera and a virtual 3D city model. The proprioceptive sensors allow to continuously estimating the dead-reckoning position and orientation of the vehicle. This dead-reckoning estimation of the pose is corrected by GPS measurements. Moreover, a 3D geographical observation is constructed to compensate the drift of the dead-reckoning localization when GPS measurements are unavailable. The 3D geographical observation is based on the matching between the virtual 3D city model and the images acquired by the camera. For that, two images have to be matched : the real image and the virtual image. The real image is acquired by the on board camera and provides the real view of the scene viewed by the vehicle. The virtual image is provided by the 3D GIS. The developed method is composed of three parts. The first part consists in detecting and matching the feature points of the real image and of the virtual image. Three methods : SURF (Speed Up Robust Features), SIFT (Scale Invariant Feature Transform) and Harris corner detector are compared. The second part concerns the position computation using POSIT (Pose from Orthographic and Scaling with Iterations) algorithm and the previously matched features set. The third part concerns the data fusion using IMM-UKF (Interacting Multiple Model- Unscented Kalman Filter). The proposed approach has been tested on a real sequence and the obtained results proved the feasibility and robustness of the proposed approach.