Abstract | ||
---|---|---|
A panoramic stereo sensor is presented which enables a Khepera robot to extract geometric landmark information ("disparity signatures") of its surroundings. By comparing the current panoramic disparity signature to memorized signatures the robot is able to return to already visited places ("homing"). Evaluating a database of panoramic stereo images recorded bythe robot we compare homing performance of the proposed disparity-based approach with a solely image-based method considering the size of catchment areas. Although the image-based technique yields larger catchment areas, disparity-based homing achieves a much higher degree of invariance with respect to illumination changes. To increase homing performance, we suggest an extended homing scheme integrating both landmark types. |
Year | DOI | Venue |
---|---|---|
2002 | 10.1007/3-540-36181-2_62 | Biologically Motivated Computer Vision |
Keywords | Field | DocType |
bythe robot,khepera robot,extended homing scheme,current panoramic disparity signature,disparity signature,disparity-based homing,catchment area,homing performance,panoramic stereo sensor,vision-based homing,panoramic stereo image | Homing (biology),Computer vision,Catchment area,Vision based,Artificial intelligence,Landmark,Robot,Geography,Stereo image | Conference |
Volume | ISSN | ISBN |
2525 | 0302-9743 | 3-540-00174-3 |
Citations | PageRank | References |
10 | 0.80 | 5 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Wolfgang Stürzl | 1 | 59 | 10.23 |
Hanspeter A. Mallot | 2 | 531 | 72.43 |