Localization of mobile robot in prior 3D LiDAR maps using stereo image sequence
The paper studies the real-time stereo image-based localization of a vehicle in a prior 3D LiDAR map. A novel localization approach for mobile ground robot, which successfully combines conventional computer vision techniques, neural network based image analysis and numerical optimization, is propose...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Samara National Research University
2024-06-01
|
| Series: | Компьютерная оптика |
| Subjects: | |
| Online Access: | https://www.computeroptics.ru/eng/KO/Annot/KO48-3/480311e.html |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832086525949509632 |
|---|---|
| author | I.V. Belkin A.A. Abramenko V.D. Bezuglyi D.A. Yudin |
| author_facet | I.V. Belkin A.A. Abramenko V.D. Bezuglyi D.A. Yudin |
| author_sort | I.V. Belkin |
| collection | DOAJ |
| description | The paper studies the real-time stereo image-based localization of a vehicle in a prior 3D LiDAR map. A novel localization approach for mobile ground robot, which successfully combines conventional computer vision techniques, neural network based image analysis and numerical optimization, is proposed. It includes matching a noisy depth image and visible point cloud based on the modified Nelder-Mead optimization method. Deep neural network for image semantic segmentation is used to eliminate dynamic obstacles. The visible point cloud is extracted using a 3D mesh map representation. The proposed approach is evaluated on the KITTI dataset and a custom dataset collected from a ClearPath Husky mobile robot. It shows a stable absolute translation error of about 0.11 – 0.13 m. and a rotation error of 0.42 – 0.62 deg. The standard deviation of the obtained absolute metrics for our method is the smallest among other state-of-the-art approaches. Thus, our approach provides more stability in the estimated pose. It is achieved primarily through the use of multiple data frames during the optimization step and dynamic obstacles elimination on depth image. The method’s performance is demonstrated on different hardware platforms, including energy-efficient Nvidia Jetson Xavier AGX. With parallel code implementation, we achieve an input stereo image processing speed of 14 frames per second on Xavier AGX. |
| format | Article |
| id | doaj-art-a31de918fa3947009e55a08fbe11b9ee |
| institution | Kabale University |
| issn | 0134-2452 2412-6179 |
| language | English |
| publishDate | 2024-06-01 |
| publisher | Samara National Research University |
| record_format | Article |
| series | Компьютерная оптика |
| spelling | doaj-art-a31de918fa3947009e55a08fbe11b9ee2025-02-06T12:18:18ZengSamara National Research UniversityКомпьютерная оптика0134-24522412-61792024-06-0148340641710.18287/2412-6179-CO-1369Localization of mobile robot in prior 3D LiDAR maps using stereo image sequenceI.V. Belkin0A.A. Abramenko1V.D. Bezuglyi2D.A. Yudin3Moscow Institute of Physics and Technology; LLC IntegrantLLC IntegrantMoscow Institute of Physics and TechnologyMoscow Institute of Physics and Technology; Artificial Intelligence Research Institute (AIRI)The paper studies the real-time stereo image-based localization of a vehicle in a prior 3D LiDAR map. A novel localization approach for mobile ground robot, which successfully combines conventional computer vision techniques, neural network based image analysis and numerical optimization, is proposed. It includes matching a noisy depth image and visible point cloud based on the modified Nelder-Mead optimization method. Deep neural network for image semantic segmentation is used to eliminate dynamic obstacles. The visible point cloud is extracted using a 3D mesh map representation. The proposed approach is evaluated on the KITTI dataset and a custom dataset collected from a ClearPath Husky mobile robot. It shows a stable absolute translation error of about 0.11 – 0.13 m. and a rotation error of 0.42 – 0.62 deg. The standard deviation of the obtained absolute metrics for our method is the smallest among other state-of-the-art approaches. Thus, our approach provides more stability in the estimated pose. It is achieved primarily through the use of multiple data frames during the optimization step and dynamic obstacles elimination on depth image. The method’s performance is demonstrated on different hardware platforms, including energy-efficient Nvidia Jetson Xavier AGX. With parallel code implementation, we achieve an input stereo image processing speed of 14 frames per second on Xavier AGX.https://www.computeroptics.ru/eng/KO/Annot/KO48-3/480311e.htmlvehicle localizationoptimizationdeep learningstereo camerasemantic segmentationembedded systems |
| spellingShingle | I.V. Belkin A.A. Abramenko V.D. Bezuglyi D.A. Yudin Localization of mobile robot in prior 3D LiDAR maps using stereo image sequence Компьютерная оптика vehicle localization optimization deep learning stereo camera semantic segmentation embedded systems |
| title | Localization of mobile robot in prior 3D LiDAR maps using stereo image sequence |
| title_full | Localization of mobile robot in prior 3D LiDAR maps using stereo image sequence |
| title_fullStr | Localization of mobile robot in prior 3D LiDAR maps using stereo image sequence |
| title_full_unstemmed | Localization of mobile robot in prior 3D LiDAR maps using stereo image sequence |
| title_short | Localization of mobile robot in prior 3D LiDAR maps using stereo image sequence |
| title_sort | localization of mobile robot in prior 3d lidar maps using stereo image sequence |
| topic | vehicle localization optimization deep learning stereo camera semantic segmentation embedded systems |
| url | https://www.computeroptics.ru/eng/KO/Annot/KO48-3/480311e.html |
| work_keys_str_mv | AT ivbelkin localizationofmobilerobotinprior3dlidarmapsusingstereoimagesequence AT aaabramenko localizationofmobilerobotinprior3dlidarmapsusingstereoimagesequence AT vdbezuglyi localizationofmobilerobotinprior3dlidarmapsusingstereoimagesequence AT dayudin localizationofmobilerobotinprior3dlidarmapsusingstereoimagesequence |