Hi colleagues,
I have been working with estimating the flight altitude of Aerial robots using 3D planar SLAM from the point cloud data of an RGB-D camera. I extract the 3D horizontal planar surfaces and map them in order to localize the aerial robot in its flight altitude.
I am starting a project where I would like to achieve also the x and y-axis localization of the aerial robots. As the RGB-D sensors are noisy, extracting the vertical planes for localizing in x and y-axis, using 3D planar SLAM techniques becomes a difficult problem.
I am looking for ideas where I can combine a 3D planar SLAM algorithm with a visual SLAM algorithm in order to achieve a robust localization in unstructured and unknown environments. I would appreciate any papers or suggestion regarding this topic.
Thanks.
Hi,
You can use the Depth information to estimate the attitude (z), and some motion detection algorithms (with image processing) to estimate the vertical location (x, y). If you intend indoor applications, you can avoid inference signals and thus reduce noise and its effects. However, for outdoor applications, I suggest using multisensor data fusion (Kalman filter for example) to have a more accurate estimation.
For indoor autonomous 2D navigation using RGB-D camera, see:
https://www.researchgate.net/publication/312234045_Optimal_path_planning_and_execution_for_mobile_robots_using_genetic_algorithm_and_adaptive_fuzzy-logic_control
Good luck,
Article Optimal path planning and execution for mobile robots using ...
3D Image processing and object recognition based on Lidar image?
Hi Friends, I want to start my work with the principle and method of 3D Image Lidar acquisition, 3D point cloud image processing, 3D Lidar image reconstruction based on the filtered 3D point cloud data, Feature extraction from the 3D reconstructed image and object recognition with simulation technique. I am looking for the Best & innovative ideas from you peoples. I would appreciate any papers or suggestion regarding this topic. Thanks
Hi Azzeddine,
Thanks for your answer. I read your paper its very interesting work. I am looking for indoors localization using point cloud sensors on-board UAVs. Since RGB-D cameras are light weight I am more interested towards using them along with small UAVs, but they have more noise as compared to 3D lidar systems.
More concretely I am looking for fusing 3D surface based SLAM with Visual SLAM in order to achieve a robust 3D localization. I keep your reasearch in mind for developing more ideas.
Thanks.
I recommend this article
https://www.mdpi.com/2504-446X/7/2/89
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Control Theory
- Control Systems