We propose the GraspBalance algorithm to achieve balanced grasp pose detection of multi-scale objects
We randomly selected one object from each of the three scales (small, medium, and large) of object collections for visualization of robot grasping.
Franka grasping in cluttered scenes consisting exclusively of small-scale objects. We randomly selected several small-scale objects placed within the robot's grasping range and visualized the grasping process.
Franka grasping in cluttered scenes consisting exclusively of medium-scale objects. We randomly selected several small-scale objects placed within the robot's grasping range and visualized the grasping process.
Franka grasping in cluttered scenes consisting exclusively of large-scale objects. We randomly selected several small-scale objects placed within the robot's grasping range and visualized the grasping process.
Real robot grasping in cluttered scenes involving objects of all scales. We randomly selected several small-scale objects placed within the robot's grasping range and visualized the grasping process.