Learning Guidance for “Bin Picking of Random Parts” Application Scenarios

Welcome to the learning guidance for “bin picking of random parts” applications.

This guidance provides you all the learning materials for implementing vision-guided robotic picking of random parts.

First, you can learn how to deploy a part picking application by referring to Bin Picking Tutorials (Master-Control Communication). Next, you can find the needed videos and documents on this page according to your robot brand, camera mounting mode, and communication mode.

If you have not yet determined the camera mounting mode and communication mode, please refer to Guidance of Docs and Tutorials.

Choose the camera model

Recommended camera model: Mech-Eye LSR L
Technical specifications: Mech-Eye LSR L

Set up the hardware of vision system

Recommended camera mounting mode: eye to hand (ETH)
Note: ETH here means that the camera is mounted on a fixed mounting frame.

Camera mounting mode Hardware setup tutorial Camera cable information
Eye to hand (ETH) How-to video: Vision System Hardware Setup
How-to guide: Vision System Hardware Setup
Camera cable features (to be added)
Guidelines for mounting camera cables (login is required)

Set up communication with the robot

Perform hand-eye calibration

Camera mounting mode
Eye to hand (ETH) Eye to eye (ETE)
Automatic calibration for 6-axis robots (Master-Control communication) How-to video: Hand-Eye Calibration in ETH Setup (Master-Control Communication)
How-to guide: Hand-Eye Calibration in ETH Setup
How-to video: coming soon
How-to guide: Hand-Eye Calibration in ETE Setup
Automatic calibration for 6-axis robots (Standard Interface communication) How-to video: Hand-Eye Calibration in ETH Setup (Standard Interface Communication)
How-to guide: Hand-Eye Calibration in ETH Setup
Manual calibration (using multiple random calibration board poses method) How-to video: coming soon
How-to guide: Manual Hand-Eye Calibration in ETH Setup

Configure vision project

Application scenario Camera parameter adjustment Vision recognition Common issues in implementation Extended requirements
Chain link picking Coming soon Coming soon A different method used to make the point cloud model Save data
Ring picking Coming soon Coming soon Gripper model in a different format More content to be added…

Perform picking

Master-Control communication Standard Interface communication Common issues in implementation
Use Mech-Viz to control robot motion Use Mech-Vision for vision recognition and get vision points from it A different communication protocol used by the robot
Use Mech-Vision for path planning and get the planned path from it (supported in V1.8.0) A different data format used by the robot
Use Mech-Viz for path planning and get the planned path from it Robot not adapted to the vision system yet
More content to be added…

Troubleshooting

Camera-related issues Vision-related issues Communication-related issues
Camera cannot be connected (login is required) Common issues in chain link picking application scenarios Summary of robot communication issues (to be added)
How to determine whether the exposure is proper (login is required) Common issues in ring picking application scenarios

Application expansions

Make the point cloud model

Picking accuracy is not required Picking accuracy should be within 0.5 mm, or the gripper dimensions are hard to know
With the CAD file of the workpiece How-to video: coming soon
How-to guide: generate the point cloud model from the CAD file and directly add a pick point in Matching Model and Pick Point Editor
How-to video: coming soon
How-to guide: generate the point cloud model from the CAD file and add the pick point by teaching in ETH setup
How-to video: coming soon
How-to guide: generate the point cloud model from the CAD file and add the pick point by teaching in EIH setup
Without the CAD file of the workpiece How-to video: coming soon
How-to guide: generate the point cloud model from the captured images and directly add the pick point
How-to video: coming soon
How-to guide: add the pick point by teaching in ETH setup and generate the point cloud model from the captured images
How-to video: coming soon
How-to guide: generate the point cloud model from the captured images and add the pick point by teaching in EIH setup

Train and fine-tune the deep learning model

Adjust the orientation and sorting of workpieces

How-to topic for pose editor (coming in V1.8.0)

Make the gripper model in OBJ format

How-to topic for model editor (coming in V1.8.0)

Set the tool and workobject symmetry

Rotational Symmetry of the Tool
Rotational Symmetry of the Workobject

Configure collision detection

Collision Detection Introduction and Configuration

Save data

How-to video: coming soon
How-to guide: Save Data