What Is This¶
Detect_graspable_poses_pcabase.py is a program which publishes pose array where a robot can grasp, using point cloud of a target object and PCA algorithm. In the above image, the grasp poses are visualized with the axies. Please transform input cloud to a flame which direction is the same as the robot (e.g. /BODY) before using this program.
The above image illustrates an example way to use this node. Axies represent poses where the robot can grasp.
Input pointcloud. RGB field is required.
PoseArray where a robot can grasp.
From which direction a robot try to grasp a target object. (’x’: from the front, ‘z’: from the top)
How long a robot can spread its hand.
Interval between target poses. Please decrease this number if you want to detect more possible grasp poses.