Hi everyone!
It's the first time I'm working with this brilliant simulator and I've found some topics very interesting like this one:
viewtopic.php?t=320
I'm using a KUKA youBot and a conveyor belt with objects moving (picture below).
My goal is to:
1 - identify those objects dimensions and orientation
2 - use youBot arm to pick them and place them in the "back plate" of the robot.
In order to achieve goal 1, I'm thinking of using a vision sensor with some filters to make the recon of the objects.
So my first question is which filters do you guys think are more appropriate for this situation, considering that the objects have different sizes, orientation, can be standing, etc....
Best regards to you all!
Picture: http://i.imgur.com/3DUhkbX.png
[First V-REP exp] Object detection and manipulation
Re: [First V-REP exp] Object detection and manipulation
Hi guys!
I hope no to be boring but I'm trying to adapt the blobDetectionWithPickAndPlace.ttt demo scene to meet my goals, but I'm having some trouble with resizing the white conveyor belt.
Is it possible to resize all parts at once? (whiteConveyor, whiteConveyor_forwarder, whiteConveyor_endCylinders and whiteConveyor_texture).
I can only resize them one by one but then I get errors with the vision sensor like this:
http://i.imgur.com/b4YGc2R.png?1
Note: all parts of the conveyor belt resize to half the original size.
Best regards to all!
I hope no to be boring but I'm trying to adapt the blobDetectionWithPickAndPlace.ttt demo scene to meet my goals, but I'm having some trouble with resizing the white conveyor belt.
Is it possible to resize all parts at once? (whiteConveyor, whiteConveyor_forwarder, whiteConveyor_endCylinders and whiteConveyor_texture).
I can only resize them one by one but then I get errors with the vision sensor like this:
http://i.imgur.com/b4YGc2R.png?1
Note: all parts of the conveyor belt resize to half the original size.
Best regards to all!
Re: [First V-REP exp] Object detection and manipulation
Hello,
Yes, you are right to use the blob detection filter, that will provide you with most information (e.g. blob size, position, orientation). Of course you can additionally probe for specific colors, etc.
To resize an object or model, use the scaling dialog in the object common properties (click the Scaling button). Make sure that only the conveyor base is selected.
If you want to scale an object differently depending on the axis, use the shape geometry dialog. In case of a model, you will have to decompose it and individually scale the objects, then recompose the model again. Grouped shapes will also have to be ungrouped in that case.
Cheers
Yes, you are right to use the blob detection filter, that will provide you with most information (e.g. blob size, position, orientation). Of course you can additionally probe for specific colors, etc.
To resize an object or model, use the scaling dialog in the object common properties (click the Scaling button). Make sure that only the conveyor base is selected.
If you want to scale an object differently depending on the axis, use the shape geometry dialog. In case of a model, you will have to decompose it and individually scale the objects, then recompose the model again. Grouped shapes will also have to be ungrouped in that case.
Cheers
Re: [First V-REP exp] Object detection and manipulation
Thank you for your extremely quick reply guys!
I have been away due to personal reasons but now I'm back to this project.
I already resized the conveyor belt and it's exactly as it's supposed.
Now the vision sensor.
I used the same as in blobDetectionWithPickAndPlace example and it works great for black objects but I want to use grey objects too, and it doesn't work the same way.
This is what I have: http://i.imgur.com/gO82DRc.png
What combination of probes do you suggest me to use to get the best results on detect grey and black objects?
Again, thank you and best regards.
I have been away due to personal reasons but now I'm back to this project.
I already resized the conveyor belt and it's exactly as it's supposed.
Now the vision sensor.
I used the same as in blobDetectionWithPickAndPlace example and it works great for black objects but I want to use grey objects too, and it doesn't work the same way.
This is what I have: http://i.imgur.com/gO82DRc.png
What combination of probes do you suggest me to use to get the best results on detect grey and black objects?
Again, thank you and best regards.
Re: [First V-REP exp] Object detection and manipulation
Hello,
try to adjust the Blob detection filter's threshold (double-click the filter to edit its threshold).
If you temporarily disable the second-last filter (Original image to work image) you will be able to see the effect of the previous filters and adjustments.
Cheers
try to adjust the Blob detection filter's threshold (double-click the filter to edit its threshold).
If you temporarily disable the second-last filter (Original image to work image) you will be able to see the effect of the previous filters and adjustments.
Cheers