In the manner of an evaluation study of stereoscopic vision systems providing partial 3D description of taken scene, we are interested in exloring its applicability to in-orbit operations as a basis for future development of vision in this field. We describe here the preliminary results. First, it is important to distinguish between stereovision display systems known in teleoperation domain which reproduce the relief impression and stereovision passive sensors providing 3D information about the environment. These two are based on the same binocular geometrical principle in which two eyes receive different images of the 3D world. The first one attempts to simulate the relief to the viewer by projecting to the left eye an image taken from the left camera of a binocular optical system, and the right image to the right eye. The viewer's perception makes he sees a 3D representation of the scene: these systems have display function. The second one, which is the object of this paper, creates 3D information by automatically matching the two views and building the 3D coordinates using a calibration model and the location of the matched points: this one has sensor function. The matching process requires pattern recognition and image interpretation techniques.
3D information from stereovision process
Verarbeitung von 3D-Informationen aus Stereovisionsverfahren
1987
5 Seiten, 10 Bilder, 11 Quellen
Aufsatz (Konferenz)
Englisch
British Library Online Contents | 2007
|Springer Verlag | 2015
|Robust homing with stereovision
SPIE | 2018
|Virtual stereovision system: new understanding on single-lens stereovision using a biprism
British Library Online Contents | 2005
|Active, Foveated, Uncalibrated Stereovision
British Library Online Contents | 2009
|