SLAM in adverse weathers: robust modality and denoised conventional modality
This dissertation aims to alleviate negative influence that adverse weathers have on SLAM system by applying filters on traditional modality and adopting more robust modality, respectively. For robust modality, the modal fusion of thermal camera, LiDAR and IMU is found out to be the most robust and...
Saved in:
| 主要作者: | |
|---|---|
| 其他作者: | |
| 格式: | Thesis-Master by Coursework |
| 語言: | English |
| 出版: |
Nanyang Technological University
2023
|
| 主題: | |
| 在線閱讀: | https://hdl.handle.net/10356/168368 |
| 標簽: |
添加標簽
沒有標簽, 成為第一個標記此記錄!
|
| 機構: | Nanyang Technological University |
| 語言: | English |
| 總結: | This dissertation aims to alleviate negative influence that adverse weathers have on SLAM system by applying filters on traditional modality and adopting more robust modality, respectively. For robust modality, the modal fusion of thermal camera, LiDAR and IMU is found out to be the most robust and effective in extreme rainy weather, and it is discovered that direct method is more suitable than indirect method for the pose estimation of thermal modality. For traditional modality, this dissertation proposes a density-based point cloud filtering method for better map reconstruction performance to detect and delete raindrops in the point cloud map, with above 90% percent of raindrops filtered out accompanied by an accuracy of above 89% in the experiment. The localization performance based on traditional modal fusion is rather enhanced by applying and setting up QR codes along the road, and experiments show great performance enhancement given different distance settings between landmarks. |
|---|