Simulation to Real Implementation on Kinova Arm
-
Install OpenCV and cv_bridge using the following commands:
$ sudo apt-get install ros-your-distro-vision-opencv $ sudo apt-get install ros-your-distro-cv-bridge -
Create a package in your workspace with the required dependencies by running the following command:
$ catkin_create_pkg package-name rospy roscpp opencv2 cv_bridge std_msgs sensor_msgs -
Put pose_estimation.py and camera_cal.py in the src folder
-
Change line 24 in camera_cal.py to the directory you have saved your camera caliberation images. (Note: Make sure to take atleast 30 images)
-
Run camera_cal.py to generate the required camera matrices. They will be saved as camera_mtx.npy and dist_mtx.npy in the same directory
-
Run roscore and your camera image publisher node
-
Change line 24 in pose_estimation.py to subscribe to your camera topic
-
Change line 38 and 39 in pose_estimation.py to the directory where camera_mtx.npy and dist_mtx.npy is saved in order to import them
-
Object pose is published on the topic /object_pose as Float32
-
Marker pose is published on the topic /marker_id as String