Scene classification from dense disparity maps in indoor environments

Darius Burschka, Gregory Hager

Research output: Contribution to journalArticlepeer-review


We present our approach for scene classification in dense disparity maps from a binocular stereo system. The classification result is used for tracking and navigation purposes. The presented system is capable of foreground- background separation classifying room structures. The 3D model of the scene is derived directly from the disparity image. This approach is used for initial target selection and scene classification in mobile navigation. It is used on our mobile system for target tracking, but can also be used for localization as described in this paper. We describe the basic principles of our object detection and classification using disparity Information from a binocular stereo system. The theoretical derivation is supported by results from the binocular stereo sensor system on our mobile robot.

Original languageEnglish (US)
Pages (from-to)708-712
Number of pages5
JournalProceedings - International Conference on Pattern Recognition
Issue number3
StatePublished - 2002

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition


Dive into the research topics of 'Scene classification from dense disparity maps in indoor environments'. Together they form a unique fingerprint.

Cite this