Welding with Robot Vision

Weld fixtures not repeatable? Inconsistent weld joints? Robots can now automatically adapt to solve these issues and more. By following these tips, you’ll see the best return from your vision investment.

The weld fixtures that took months of development time and were purchased for tens of thousands of dollars are not repeatable. Rework costs are going to be astronomical. The thought of starting production like this can make you ill.

One solution to these inconsistencies is robot vision. Giving your robot the ability to automatically adapt to inconsistent parts and variable weld joints can save the day. Proactive robot manufacturers have meticulously developed their own vision hardware and software that fully integrates into the robot controller. All you do is plug a camera into the robot control cabinet and – wha-lah! – you have vision (after you pay for the option, of course).

Fully integrated robot vision is a powerful tool, easy to set up with a lot of capability to locate and track almost every type of weld joint. Robot vision is not just limited to tracking weld joints, of course. It can also check for part presence, measure characteristics such as hole diameter, and perform many other quality-oriented functions. If you are worried about how difficult it is to use, don’t be. On average, it takes about three days to learn and become fully comfortable with how an integrated vision system works.

2D OR 3D
Once you have decided to implement a vision system into your work cell, the next important decision should be whether to use a 2D or 3D camera. If you remember your 7^th grade geometry class and the X, Y, Z coordinate system, 2D uses the X and Y coordinates and 3D uses X, Y and Z. In other words, X is forward and back, Y is left and right, and Z is up and down.

If your weld joint is moving in the X and Y directions (2D), then you can save some cash and purchase a 2D camera system. However, if your weld joint is moving X, Y and Z (3D), then you will have to purchase a 3D camera that uses a type of laser system to track and adjust for the height change. 3D systems cost a little more but do a fantastic job at adjusting for height changes and can even account for yaw and skew of the part.

CONTRAST & CONSISTENCY
Robotic vision is all about contrasts and the consistency of that contrast. For example, take the contrast of black objects vs. white ones. Most vision systems “see” in shades of gray. When one pixel has sufficient light to dark difference as compared to the one next to it, it triggers an edge line. The unique image created by these edge lines is the basic principle used by robot vision.

Contrast on most objects this is accomplished by using carefully selected lighting or lasers. By controlling the angles of lighting and camera position, almost everything welded can use vision to give weld joint position information to the robot, even curved surfaces.