DoF - a Robotiq Community
Warning sign
The Dof Community was shut down in June 2023. This is a read-only archive.
If you have questions about Robotiq products please reach our support team.
Categories Discussions
Home > Programming > Increasing the Wrist Camera's Field of View
Tyler_Berryman
Enric

@Tyler_Berryman

I have been testing this function. First of all I attach the video for you to see the process.

I want to recognize the face of the dice.
First we recognize the dice (parametric method) and then the number of points (automatic method).

The main problem is that I have tried to enlarge the image to have a certain percentage between if there is a "one point" or there is no "one point".

The dice is yellow and transparent ... It does not work. If the dice were opaque it works well. (in the video).

Do you have any idea to make the process more robust?

Thank you! and see you soon!

https://www.youtube.com/watch?v=xt_BIlazbqI
David_Gouffe

Hi Enric, 

It is an old post but it still can be useful to some people. 

To improve the success, using a bigger dice would be the prime approach. The rule of thumb for object size to be detected by the camera is 1/10 of the field of view. Minimum being 10 mm x 7.5 mm, object smaller than 10 mm x 7.5 mm is subject to be undetected. Note that in this case, the work distance is 7 cm from the surface to look at. 

Transparent part can be a challenge too. 

With the newest URCAP for the camera it is possible to search for multiple time the same object (ex.: the dot on the die). It is also possible to use advance mode to control advanced parameters to target the most contrasting parameters. 

If it is still not working after that you can improve the lighting of the scene and block side light but this is just for challenging vision applications or hard to find parts.

If you want to look at the 6 side of the die and match each one without counting the number of dot: Teach the most restrictive face (6 dot) then reduce the number of dots searched/taught in automatic mode. One face at a time up to the last, face with a single dot. This produce a bigger program but it is more reliable as long as the die model stay the same. The point here is that a single dot can be found in the 5 and 3 dots faces. 

Using the camera is the best way to know its limits. And now those limits can be quite far with all the advanced control. 

Regards,