Depalletizing Skill

Overview

This skill is used for depalletizing single-variety crate types. The box size (length and width) has to be specified in the robobrain® system. Using AI, the boxes are to be reliably recognized and the gripping points in the center of the box are returned to the robot for each box within the visible layer. The skill is implemented such that the box layers are processed from top to bottom to ensure that no boxes of the lower layer are grabbed before the upper layer is completely emptied. Specifying a workspace (ROI) ensures that the skill only focuses on the area of crates and examines them for grab points.

The input to the skill is color and depth information, which is processed by the AI skill. This information is used to identify the different box layers and the box dimensions within them. By specifying the box dimensions in the robobrain® system's graphical interface, these can be reliably identified. In the event of changing crate sizes, the new dimensions can be adjusted in the GUI.

The returned gripping points per crate always represent the center of the box. This is returned next to the quality value and displayed on the graphical interface.

Skill Result Information

Position	Position of the grasp point relative to the robot’s coordinate system (in m)
Orientation	Rotation of the grasp point relative to the robot’s coordinate system (in m)

Specifications

Conditions

Camera mount:

• Dynamic

• Static

Camera distance:

1,25 – 3,5 m

Parts dimensions:

20 – 40 cm per side

Setup:

Camera position is central above the pallet and camera parallel to the pallet.

Specs

Avg. recognition time:

1 seconds (480x640px)

4 seconds (1920×1080px)

Supported grippers:

• Vacuum

Features

• Detection of different layers

• Detection of all boxes of a layer

Technical Parameter Description

Parameter

Basic

Name

The width in [m] of the boxes that shall be detected

Parameter

box_width

Type

flaot

Description

In order for the skill to robustly detect the boxes, the width of the box to be detected must be entered here in meters.

Name

The height in [m] of the boxes that shall be detected

Parameter

box_height

Type

float

Description

In order for the skill to robustly detect the boxes, the height of the box to be detected must be entered here in meters.

Expert

Name

​Optimize gripper orientation

Parameter

orientation_optimization

Type

boolean

Description

Enables the optimization of the gripper orientation. If enabled the pick pose might be rotated by 180° around the approach axis to minimize the robot rotation between the capture and pick pose.

Detections

Detection Parameter

Type

pose (Transformation)

Description

Pick pose for the gripper

Type

data (int)

Description

Quality of the original detection, as reported by the AI model.

Type

optimization_angle (float)

Description

Only present if the parameter optimization_orientation is set to true. The parameter represents the rotation angle in radians that was applied to the original pick pose during the orientation-optimization step.

Detection Example

{
  "detections":[
    {
      "pose":{
        "x":0.4400900392852657,
        "y":-0.20171213837105148,
        "z":-0.5735028870670782,
        "rx":-0.0005108216318221825,
        "ry":-3.1372998612103395,
        "rz":-0.03927747019701938
      },
      "data":"11",
      "optimization_angle":3.141592653589793
    },
    {
      "pose":{
        "x":1.0319483363064892,
        "y":0.1901773268982383,
        "z":-0.5650816200514848,
        "rx":-0.0005108216318221825,
        "ry":-3.1372998612103395,
        "rz":-0.03927747019701938
      },
      "data":"4",
      "optimization_angle":3.141592653589793
    },
    {
      "pose":{
        "x":1.0323984722672992,
        "y":-0.19841698316530285,
        "z":-0.5628114599802168,
        "rx":-0.0005108216318221825,
        "ry":-3.1372998612103395,
        "rz":-0.03927747019701938
      },
      "data":"10",
      "optimization_angle":3.141592653589793
    }
  ]
}