| Meng Xujiong（孟旭炯）,Jiang Rongxin,Chen Yaowu.[J].高技术通讯(英文),2012,18(3):238~242 |
|
| Modified homogeneous parameterization for monocular SLAM |
| |
| DOI：10.3772/j.issn.1006-6748.2012.03.004 |
| 中文关键词: |
| 英文关键词: monocular simultaneous localization and mapping (SLAM), feature initialization, depth estimation, homogeneous coordinates, extended Kalman filter |
| 基金项目: |
| Author Name | Affiliation | | Meng Xujiong（孟旭炯） | | | Jiang Rongxin | | | Chen Yaowu | |
|
| Hits: 1353 |
| Download times: 2008 |
| 中文摘要: |
| |
| 英文摘要: |
| Feature initialization is an important issue in the monocular simultaneous localization and mapping (SLAM) literature as the feature depth can not be obtained at one observation. In this paper, we present a new feature initialization method named modified homogeneous parameterization (MHP), which allows undelayed initialization with scale invariant representation of point features located at various depths. The linearization error of the measurement equation is quantified using a depth estimation model and the feature initialization process is described. In order to verify the performance of the proposed method, the simulation is carried out. Results show that with the proposed method, the SLAM algorithm can achieve better consistency as compared with the existing inverse depth parameterization (IDP) method. |
|
View Full Text
View/Add Comment Download reader |
| Close |
|
|
|
