Real-Time Autonomous System of Navigation Using a Stereoscopic Camera

Artificial intelligence is in a continuous development and automation of vehicles is a discussed subject. The navigation of any vehicle is based on obstacle avoidance, but it still challenges the field of automation because of the unpredictable environment. Usually, an algorithm for obstacle avoidan...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:2018 International Conference on Communications (COMM) S. 497 - 500
Hauptverfasser: Mihalcea, Georgiana, Suciu, George, Vasilescu, Cristian
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 01.06.2018
Schlagworte:
Calibration
Cameras
Collision avoidance
disparity map
Estimation
Image coding
LHE
Microsoft Windows
obstacle avoidance algorithm
Sensors
stereoscopic camera
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Artificial intelligence is in a continuous development and automation of vehicles is a discussed subject. The navigation of any vehicle is based on obstacle avoidance, but it still challenges the field of automation because of the unpredictable environment. Usually, an algorithm for obstacle avoidance is based on several types of sensors. This article will present an evaluation of two algorithms, one using the mean estimation method and one using the threshold estimation method, both based on only one imaging sensor, which simplify the complexity of the system. The algorithms are based on 3D reconstruction of the environment and making the disparity map which is a matrix that contains the depth information of each pixel from an image. Both algorithms have been implemented in C++, using as device DUO MLX stereo camera on an embedded system with ARM processing unit (DUO VPC).
DOI:10.1109/ICComm.2018.8484777