How to identify 2 different parts with the camera in the same program?

Catherine_Bernier

I got a question from a customer about if we can teach 2 different parts in the same program. The answer is yes you can! Just like described in this post, you can create a camera node to detect the part, then use the camera again to detect the tray or box to put the part in. Moreover, you don't need to calibrate the snapshot position at the same height for both. You can have different positions for each calibration.

I'll post a video later this week describing the process.

Any applications in mind where you need such a feature?

Tyler_Berryman

@Catherine_Bernier Like you mentioned you can have several different Camera Locate nodes in the same program. You can use several Camera Locate nodes to look for 2 different parts in the same program. Another possible application is using multiple Camera Locate nodes to create a larger field of view, this can be done by changing the snapshot position for each Camera Locate node. Another application consists of looking for parts that have very different sizes, you can use one camera locate node to look for a part that is 7"x7" and a second Camera Locate node to look for a smaller part (2"x2") that needs to be assembled onto the larger part.

Catherine_Bernier

Exactly @Tyler_Berryman! Here is the video I talked about and its explanation.

So in this example, I use a first snapshot position to detect the overall part on the workstation. Then the robot picks up the part and moves it to another location where a closer snapshot position is used. This is done in order to get a more precise view on the part and detect details like the Robotiq logo in this video. 

When creating another Camera Locate node, you are asked to choose the right snapshot position. Note that a calibration is always required when creating a snapshot position. You can see a detailed video on how to define your snapshot position and calibrate the camera here.

Also, as you will see in the video, the snapshot position and the calibration position can be different. In my example, I had to position the camera near to the table in order to see the part precisely. This meant I had to define the snapshot position at this height. When it was time to calibrate, I had to position the robot to have a complete view of the calibration board. Once the calibration is done, the robot will always use the snapshot position to recognize the part.

During the teaching process, the robot will have to move and take 9 poses of the part. Do not worry if you are near to the table since the 9 poses are taken from the calibration position which is higher.

Finally here is the program I used:

Tim_Smith

@Catherine_Bernier So this means that the snapshot position and the calibration position can be different?

And does this mean that I can use the Robotiq camera to complete some simple inspection tasks? All I want to do is verify that a certain component is present following an assembly process.

Catherine_Bernier

@Tim_Smith Yes, exactly. So the snapshot position and its calibration position don't need to be the same. If you want to locate a specific small feature on a part, you can simply have a snapshot position nearer from the part.