Optical flow based navigation

OData support
Dr. Loványi István
Department of Control Engineering and Information Technology

Using cameras fixed on a vehicle we can implement many functions, which help the driver on the road and increase the vehicle’s safety. These functions, systems are for example pedestrian detection, path tracking, keeping distance, auto-parking program, navigation and mapping an unknown environment. The great advantage of these systems is that carry out tasks without human intervention, and often better than people expect it to. Vehicles of the near future quite probably will contain many of these functionalities. In this thesis I present algorithms, which I implemented in a program calculating the trajectory of a moving camera, using only the image sequence output of this sensor. The program provides the relative movement of the observer (camera) related to the environment.

3D navigation, in other word the estimation of the 2D camera movement and at the same time the mapping of the unknown environment are based on the detection and tracking of a large number of feature or template like image data.


Please sign in to download the files of this thesis.