Abstract
Abstract
Computer vision is an interesting and fast growing topic and deals with the 3D-reconstruction
of natural pictures. There are a lot of applications, research papers, algorithms and programs for
this topic available. In computer vision stereo-camerasystems are used. Depending on the eld of
application (close-up range or aerial photograph) di erent considerations have to be taken.
In the company arsenal research the road-measuring vehicle RoadSTAR was enhanced with two
digital video cameras, which record the front view of the street. In the postprocessing stereo
pictures can be extracted from the videos. In these stereo pictures distance measurements of
objects (e.g. height of a bridge or width of a road) in a distance up to 30 m should be possible.
Within this master thesis the basic theory of measuring in stereopictures is developed. In the
practical execution the camera calibration, stereo calibration, a few factors on the accuracy,
the measurement software and the rst results of distance measurements are described. In the
conclusion the usability of the stereo-camerasystem is assessed and it is discussed which further
requirements have to be ful lled to create a system, which can be used for distance measurements
in a close-up range.
Original language | German |
---|---|
Awarding Institution |
|
Supervisors/Advisors |
|
Award date | 31 Aug 2010 |
Publication status | Published - 2008 |
Research Field
- Former Research Field - Mobility Systems
- Former Research Field - Digital Safety and Security
Keywords
- Keywords