MULTI SENSOR BASED OBJECT DETECTION IN DRIVING SCENES

Xu, Philippe

Description

Niveau: Supérieur, Master

• rapport de stage

MULTI-SENSOR BASED OBJECT DETECTION IN DRIVING SCENES Internship report, Research Master's Degree in Computer Science ENS Cachan/Rennes and University of Rennes 1 Internship dates : February-June 2011 At Peking University, LIAMA, Beijing, China Key Laboratory on Machine Perception PHILIPPE XU ENS Cachan/Rennes University of Rennes 1 Advisors : FRANCK DAVOINE HUIJING ZHAO Researcher, CNRS Professor, Peking University LIAMA, Sino-French Laboratory Key Lab on Machine Perception Abstract The work done in this internship consists in two main part. The first part is the design of an experimental platform to acquire data for testing and training. To design the experiments, onboard and onroad sensors have been considered. A calibration process has been conducted in order to integrated all the data from different sources. The second part was the use of a stereo system and a laser scanner to extract the free navigable space and to detect obstacles. This has been conducted through the use of an occupancy grid map representation. Keywords: Stereo camera, Sensors calibration, Occupancy grid map, Obstacle detection June 2, 2011 du m as -0 06 36 82 2, v er sio n 1 - 2 8 O ct 2 01 1

• calibration


• obstacles like


• has been


• camera stereo


• obstacle


• stereo based


• been conducted through

Sujets

• Peking University
• Calibration
• Has Been
• Obstacle

Informations

• Publié par: dumas_ccsd
• Nombre de lectures: 20
• Langue: English
• Poids de l'ouvrage: 9 Mo

Extrait