search

mcgill-robotics / rover

Mars Rover Project software development
GNU General Public License v2.0
13 stars 8 forks source link

Obstacle detection with IntelRealsense Depth and Tracking Cameras (AUTONOMY) #123

Open amotaouakkil123 opened 8 months ago

amotaouakkil123 commented 8 months ago

Branch: automation/object-detection-and-bounding-boxes

Need to interface with IntelRealSense Depth Camera and map real world objects with this object detector.

Tech Dependencies: x64/AMD64 computer, Intel Realsense Depth Camera and its SDK. Task Dependencies: None

  1. Set up Realsense Developer Kit.
  2. Download the drivers for ROS.
  3. Set up autonomy branch and test the simulation.
  4. Implement a Realsense publishers that allow for real-time access to various data from Depth and Tracking cameras.
  5. Implement the publishers within the simulation.
  6. Test and validate through simulation and Pull Request.
  7. Merge into main.
amotaouakkil123 commented 8 months ago

ROS Drivers for IntelRealsense Depth and Tracking: https://github.com/IntelRealSense/realsense-ros/tree/ros1-legacy

Intel Info on Depth Cam: https://www.intelrealsense.com/get-started-depth-camera/

Intel Info on Tracking Cam: https://www.intelrealsense.com/tracking-camera-t265/

amotaouakkil123 commented 8 months ago

https://github.com/IntelRealSense/realsense-ros/wiki/SLAM-with-D435i

Potential map already created from open source intel realsense library.

https://github.com/IntelRealSense/realsense-ros/tree/ros1-legacy

Above link contains a lot of useful tools for mapping.

amotaouakkil123 commented 6 months ago

Realsense libraries were integrated. Rtabmap is not well integrated in realsenseSDK. Will need to switch to fitting shapes from depth map.