Vision-Guided Single-Case Depalletizing
This tutorial introduces how to deploy a 3D vision–guided carton depalletizing application using the application template case of “Single-Case Depalletizing” in the Solution Library.
Application scenario: The 3D vision system guides the robot to pick single-case cartons from the pallet and place them on the conveyor line.
Application Overview
-
Target object: single-case cartons.
-
This application uses a real camera to capture images of cartons for target object recognition. If you want to use a virtual camera, please click here to download image data of the cartons.
-
This application uses deep learning to assist recognition. There is already a built-in deep learning model package in the
resource/dl_modeldirectory of this solution.
-
-
Camera: Mech-Eye DEEP camera, mounted in eye to hand (ETH) mode.
-
Calibration board: It is recommended to use the calibration board CGB-050.
-
Robot: a six-axis robot. This application uses ABB_IRB6700_150_3_20 as an example.
-
IPC: Mech-Mind IPC STD
-
Software: Mech-Vision & Mech-Viz 2.0.0, Mech-Eye Viewer 2.4.0
-
Communication solution: Standard Interface communication, in which the vision system outputs the path planned by the Mech-Viz software.
-
End tool: vacuum gripper
For this application, you are required to prepare a model file in .obj format for the vacuum gripper, which will be used for collision detection during path planning. You can download it by clicking here.
-
Scene object: scene object model
This application requires a scene model file in .stl format, which is used to simulate a real scene and is used for collision detection in path planning. You can download it by clicking here.
|
If you are using a different camera model, robot brand, or target object than in this example, please refer to the reference information provided in the corresponding steps to make adjustments. |
Deploy a Vision-Guided Robotic Application
The deployment of the vision-guided robotic application can be divided into six phases, as shown in the figure below:
The following table describes the six phases of deploying a vision-guided robotic application.
| No. | Phase | Description |
|---|---|---|
1 |
Vision Solution Design |
Select the hardware model according to the project requirements, determine the mounting mode, vision processing method, etc. (This tutorial has a corresponding vision solution, and you do not need to design it yourself.) |
2 |
Install and connect hardware of the Mech-Mind Vision System. |
|
3 |
Load the robot master-control program and the configuration files to the robot system and set up the communication between the vision system and the robot, thus helping the Mech-Mind Vision System obtain control over the robot. |
|
4 |
Perform the automatic hand-eye calibration in the eye-to-hand setup, to establish the transformation relationship between the camera reference frame and the robot reference frame. |
|
5 |
Use the application template “Single-Case Depalletizing” in Mech-Vision Solution Library and plan the robot path with the advanced component of path planning. |
|
6 |
Based on the robot example program MM_S9_Viz_RunInAdvance, write a pick-and-place program suitable for on-site applications. |
Next, follow subsequent sections to complete the application deployment.