Abstract | ||
---|---|---|
Autonomous navigation and localization remains a challenging task for a moving object. Computer-vision based solutions have been proven effective when estimating the location of a moving object from an on-board camera using feature points. These feature points can be created in the image where the gradients change in all directions and offer unique values. By observing the position changes of feature points, the motion of the camera can be determined. As a result, feature detection and matching play critical role when estimating the motion of camera using video. In this paper, a GPU-accelerated feature detecting, and matching motion estimation system is introduced and tested using a drone platform. During our test, the GPU-accelerated system showed a 4-times increase in speed when compared with CPU based approach. |
Year | DOI | Venue |
---|---|---|
2019 | 10.1109/EIT.2019.8834204 | 2019 IEEE International Conference on Electro Information Technology (EIT) |
Keywords | Field | DocType |
SLAM,GPU acceleration,CUDA | Computer vision,Feature detection,CUDA,Computer science,Electronic engineering,Design flow,Artificial intelligence,Drone,Motion estimation | Conference |
ISSN | ISBN | Citations |
2154-0357 | 978-1-7281-0928-2 | 0 |
PageRank | References | Authors |
0.34 | 10 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Guojun Yang | 1 | 4 | 2.83 |
Jafar Saniie | 2 | 152 | 47.55 |