Current Trends in Robotic Vision and Guidance

There have been some interesting developments in both robotic guidance and machine vision over the past couple years that is deserving of some discussion.

Firstly, I would like to point out that there are a lot of resources out there to plan and understand machine vision inspection and more specifically, lighting for these systems and the types of filters to be used.

This will not be that type of resource. I will instead be focusing on the types of vision that have been of interest recently, more specifically the guidance of either the robots application onto an object, or the finding and interacting (pick and place) with parts or objects.

This list will not be all-inclusive. I will discuss the systems and solutions that I have worked with. What worked well, and what did not. Each system has its own benefits and drawbacks.

One of the most important things I would like to share is how the systems function, which may explain some of the reasons why they excel in some areas, but not in others.

Manufacturer Specific – Manufacturer specific tools have many advantages such as the ability to integrate easily with supplied hardware and a native interface. Generally, the vision application is well documented, and because image processing is done internally, external communication setup is generally not necessary, but possible.

iR Vision (Fanuc) – This is divided up into 2 subcategories:

2D Vision – Specialize in pick and place. Allows for variations in the pick or set position based on the difference between taught and recorded position. Lower cost only utilizes light and a single camera

3D Vision – This can be accomplished with a laser and a camera, multiple cameras or and area scanner that plots data points for all objects. The laser is the most common and is generally mounted on the end effector to enable the robot to adjust the position at the part pickup.

Motosight (Motoman) – Has both 2D and 3D applications each have their respective equipment and setup. These systems rely heavily on Cognex cameras and integrated interface similar to the In-Sight Explorer software.

2D Vision – Allows for variations in the pick or set position based on the difference between taught and recorded position. Utilizes a single camera, generally a fixed camera.

3D Vision – This can be accomplished with a laser and a camera, multiple cameras or and area scanner that plots data points for all objects. The laser is the most common and is generally mounted on the end effector to enable the robot to adjust the position at the part pickup.

 

Camera Vendor Solutions – These solutions are able to be used with multiple robot manufacturers, but may not have specific documentation for your setup. These types are most commonly used for part verification and error proofing (Robot shows part to camera). 

Robeye – This type of vision system is more dynamic because the robot takes multiple images from different angles in order to find the part position in the 3d space, using only one camera. This system allows for the integration of multiple robot manufacturers because the positional math is done in the robot, and the robot interfaces through ethernet with the standalone control panel. Robot Carried camera is the most common type of vision of this system.

The most common use of these last two systems is error proofing or part dimensioning confirmation (Quality). Ths use of this system with non- native hardware causes the need for calculation of the offset to be done in the PLC and requires extensive knowledge of measurement tools and transferring the data back into the robot. (data types and interpretation of the values on the robot side.)

Cognex – This system is generally combined with the manufacturer software (In-Sight Explorer) to provide error proofing and verification of important features of a build process (identifying bolts, labels, barcodes or other important components in a process). This process can be used for x and y position guidance, but still proves to be complicated to move data into the robot for offsetting robot position.

Keyence – This system is generally combined with the manufacturer software (CV-X series) This system has shown good results for identifying critical component presence or dimensions. I have not seen an easy to integrate robot guidance solution. Data retained must be processed through the PLC and translate to the robot through IO.

Omron – This system is generally combined with the manufacturer software (FJ Series) This system has shown good results for identifying critical component presence or dimensions. I have not seen an easy to integrate robot guidance solution. Data retained must be processed through the PLC and translate to the robot through IO.

 

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s