Timetable : 26 hours (6h C, 8h TD, 12h TP)

Teachers: Samia Boukir, Bernard Besserer

Requisites : Image Analysis (U.E. LGI 8224), Statistical Data Analysis (U.E. LGI 8232). Good knowledge of C language and Maths.

Purposes : This course describes first, existing 2D and 3D sensors and then, deals with the recovery of 3D information using a single or several cameras.

Contents : 2D sensors (video signal characteristics and CCD sensors). 3D sensors (principles of triangulation and laser/ultrasonic range finder). Extraction of 2D primitives (Hough transform, least squares estimation). Matching 2D primitives over an image sequence (correlation, relaxation, hypothesis prediction and verification). 3D Reconstruction (stereovision, dynamic vision).

Assessment : One term exam (E) and a continuous assessment (CA) - U.E. assessment = (E+CA)/2

References :

1. Vision 3D. Analyse d'images pour la maîtrise de génie informatique. Polycopié, S.Boukir, septembre 1996.
2. Vision stéréoscopique et perception multisensorielle. Applications à la robotique mobile. N.Ayache, Collection Science Informatique, InterEditions, Paris, 1989.
3. Vision . A computational investigation into the human representation and processing of visual information. D.Marr, W.H.Freeman and Compagny, 1982.