husky_tidy_bot_cv

Installation:

• Install dependencies:

# subprocess-tee
cd ~
git clone http://github.com/andrey1908/subprocess-tee
cd ~/subprocess-tee
pip install .

• Build from container:

source /opt/ros/noetic/setup.bash 
cd tracking/skillbot_3d_tracking
catkin_make_isolated -DCMAKE_BUILD_TYPE=Release

---

Running:

Download objects models to ~/tracking/skillbot_3d_tracking/src/husky_tidy_bot_cv/objects_models.

Optional: To run YOLOv8 segmentation for toy boxes, download model config and weights to ~/tracking/skillbot_3d_tracking/src/husky_tidy_bot_cv/yolov8_weights folder.

To run nodes use corresponding scripts from bin folder.

0. Change cameras configuration:

bash bin/config_cameras.sh

1. Segmentation YOLOv8 (for testing purposes)

bash bin/yolov8_node.sh

2. 2D-Tracking

bash bin/bot_sort_node.sh

3. Point Cloud extraction

bash bin/object_point_cloud_extraction_node.sh

4. 3D-Tracking (needs cartographer to work)

bash bin/tracker_3d_node.sh

5. Object pose estimation

bash bin/object_pose_estimation_node.sh