ICFICE 2018
Current Result Document :
| 한글제목(Korean Title) |
Stereo Visual Localization, Sending and Perception for Fast-Moving Unmanned Ground Vehicles |
| 영문제목(English Title) |
Stereo Visual Localization, Sending and Perception for Fast-Moving Unmanned Ground Vehicles |
| 저자(Author) |
Han Wang
Rui Jiang
Handuo Zhang
Karunasekera Hasith
Hui Zhou
|
| 원문수록처(Citation) |
VOL 10 NO. 01 PP. 0343 ~ 0347 (2018. 06) |
한글내용
(Korean Abstract) |
|
영문내용
(English Abstract) |
This paper presents a stereo vision system for unmanned ground vehicle navigation, where road feature detection, obstacle detection, and visual SLAM modules have been developed on two pairs of stereo cameras. To fully utilize characteristics of features at different distances, hybrid features are parametrized in visual SLAM. An absolute heading is used to reduce rotation and translation drift of Visual Odometry (VO). Negative obstacle detection has been proposed to deal with complex road conditions. Road features including lanes and curbs are detected by using distance transform on the intensity image and through disparity map, respectively. Compared to the conventional lidar-based system, the proposed solution enables low-cost implementation of localization, obstacle detection and road feature detection for autonomous vehicles.
|
| 키워드(Keyword) |
Unmanned ground vehicle (UGV)
localization
sensing
stereo vision
|
| 파일첨부 |
PDF 다운로드
|
