uptopia/affordance_ros

download weighting for YOLOv4, DeepLab V3+

YOLOv4 weight link put files in folder workspace/src/affordance_ros/part_detection/weights/20210806/

DeepLab V3+ weight link put files in folder workspace/src/affordance_ros/part_sematic_seg/weights/

#YOLOv4
chomod +x demo_d435i.py

. devel/setup.bash
roslaunch part_detection part_detection.launch

#DeepLab V3+
chomod +x juice_seg.py
chomod +x popcorn_seg.py
chomod +x show_seg_all.py

. devel/setup.bash
rosrun part_sematic_seg juice_seg.py
rosrun part_sematic_seg popcorn_seg.py
rosrun part_sematic_seg show_seg_all.py

Dependencies

• Install realsense-ros
• Install cv_bridge for Python3

Software environment

• Ubuntu 18.04
• ROS Melodic
• Python 3.6.9
• opencv 4.5.1 cv2 (?)
• cv_bridge (python3 待測試) (?)
• Install pcl (?)
• Install cv_bridge(?)

1. 部件分割 affordance_ros

• (1) part_detection: YOLOv4, YOLOv4-tiny - popcorn_f, popcorn_b, juice
  
  part_detection_result

• (2) part_sematic_seg: DeepLab V3+ - plant, handle, cap, body
  
  part_sematic_seg(juice), part_sematic_seg(popcorn)

  

part_sematic_seg(juice,popcorn)