Author | Conference | Journal | Organization | Year | DOI
Look for results that meet for the following criteria:
Academic
Publications
Realtime Omnidirectional Stereo for Obstacle Detection and Tracking in Dynamic Environments

Realtime Omnidirectional Stereo for Obstacle Detection and Tracking in Dynamic Environments,Hiroshi Koyasu,Jun Miura,Yoshiaki Shirai

Edit
Realtime Omnidirectional Stereo for Obstacle Detection and Tracking in Dynamic Environments   (Citations: 20)
BibTex | RIS | RefWorks Download
This paper describes a realtime omnidirectional stereo system and its application to obstacle detection and track- ing for a mobile robot. The stereo system uses two omnidi- rectional cameras aligned vertically. The images from the cameras are converted into panoramic images, which are then examined for stereo matching along vertical epipolar lines. A PC cluster system composed of 6 PCs can gen- erate omnidirectional range data of 720x100 pixels with disparity range of 80 about 5 frames per second. For ob- stacle detection, a map of static obstacles is first gener- ated. Then candidates for moving obstacles are extracted by comparing the current observation with the map. The temporal correspondence between the candidates are es- tablished based on their estimated position and velocity which are calculated using a Kalman filter-based tracking. Experimental results for a real scene are described. This paper describes a realtime omnidirectional stereo system and its application to detection of static and dy- namic obstacles. The system uses a pair of vertically- aligned omnidirectional cameras. The input images are converted to panoramic images, in which epipolar lines be- come vertical and in parallel. The stereo matching is per- formed by a PC cluster system to realize a realtime range calculation for a relatively large image size. For obstacle detection, a map of static obstacles is first generated. Then candidates for moving obstacles are extracted by compar- ing the current observation with the map. The tempo- ral correspondence between the candidates are examined based on their estimated position and velocity which are calculated using a Kalman filter-based tracking.
Conference: International Conference on Intelligent Robots and Systems - IROS , 2001
Cumulative Annual
Sort by:

Citations (20)

Share this on