OCCRA
Go to Post Chairman's Award is not about building the robot. Every team builds a robot. - Retired Starman [more]
Home
Go Back   Chief Delphi > Technical > Technical Discussion
CD-Media  
portal register members calendar search Today's Posts Mark Forums Read FAQ rules

 
Reply
Thread Tools Rate Thread Display Modes
  #61   Spotlight this post!  
Unread 12-19-2017, 08:30 AM
gerthworm's Avatar
gerthworm gerthworm is offline
Making the 1's and 0's
FRC #1736 (Robot Casserole)
Team Role: Mentor
 
Join Date: Jan 2015
Rookie Year: 2015
Location: Peoria, IL
Posts: 706
gerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond repute
Re: JeVois Smart Machine Vision

Quote:
Originally Posted by AllenGregoryIV View Post
Has anyone figured out the best way to talk to the JeVois when it identifies as /dev/ttyACM0?
I won't claim this is the best way, but we've been having luck so far using the built-in WPILib functions, and have done some abstraction for running commands, checking for OK with timeout, and parsing for packets delimited by {}. Seems to be working, at least.

https://github.com/RobotCasserole173...Interface.java
Reply With Quote
  #62   Spotlight this post!  
Unread 12-19-2017, 09:14 AM
billbo911's Avatar
billbo911 billbo911 is offline
I prefer you give a perfect effort.
AKA: That's "Mr. Bill"
FRC #2073 (EagleForce)
Team Role: Mentor
 
Join Date: Mar 2005
Rookie Year: 2005
Location: Elk Grove, Ca.
Posts: 2,662
billbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond repute
Re: JeVois Smart Machine Vision

Just as an FYI, Anand ("Asid61") has posted a White Paper on the use of GRIP with JeVois in FRC..
__________________
CalGames 2009 Autonomous Champion Award
2012 Sacramento Semi-Finals, 2012 Sacramento Innovation in Control Award, 2012 SVR Judges Award.
2012 CalGames Autonomous Challenge Award winner ($$$).
2014 2X Rockwell Automation: Innovation in Control Award (CVR and SAC).
Curie Division Gracious Professionalism Award.
2014 Capital City Classic Winner AND Runner Up. Madtown Throwdown: Runner up.
2015 Innovation in Control Award, Sacramento.
2016 Chezy Champs Finalist, 2016 MTTD Finalist
2017 Utah Regional Winner!, Sacramento Finalist
Innovation in Control Newton/Carver Divisions , Newton #5 Captain
2018 WFFA Sacramento, Creativity Award Galileo-Robling Divisions
Reply With Quote
  #63   Spotlight this post!  
Unread 12-19-2017, 12:57 PM
Cog's Avatar
Cog Cog is online now
Predictor and Neatifier
AKA: Cam Ogden
FRC #3324 (The MetroBots)
Team Role: Tactician
 
Join Date: Feb 2016
Rookie Year: 2010
Location: Ohio
Posts: 118
Cog has a brilliant futureCog has a brilliant futureCog has a brilliant futureCog has a brilliant futureCog has a brilliant futureCog has a brilliant futureCog has a brilliant futureCog has a brilliant futureCog has a brilliant futureCog has a brilliant futureCog has a brilliant future
Re: JeVois Smart Machine Vision

So as someone who has only tangential experience programming, can anyone sum up for me the differences between Pixy and JeVois?

Both seem pretty similar at first glance, being small compact cameras with on-board vision processing.
__________________
Alliances win matches, not robots.

Unless you're Team Hammond in 2002
Reply With Quote
  #64   Spotlight this post!  
Unread 12-19-2017, 01:11 PM
marshall's Avatar
marshall marshall is offline
Online Stripe & Drape Sales
FRC #0900 (The Zebracorns)
Team Role: Mentor
 
Join Date: Jan 2012
Rookie Year: 2003
Location: North Carolina
Posts: 3,018
marshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond repute
Re: JeVois Smart Machine Vision

Quote:
Originally Posted by Cog View Post
So as someone who has only tangential experience programming, can anyone sum up for me the differences between Pixy and JeVois?

Both seem pretty similar at first glance, being small compact cameras with on-board vision processing.
I'm going to use an analogy - it's not a perfect analogy but it's what I got so bear with me.

In the world of FRC machine vision you have options similar to the world of cars:

You can go the Honda Fit route with something that will work and a lot of people use and like and is well supported: The USB camera and RoboRIO and/or the Axis camera. Though I think the Axis is likely more Ford Focus but like not the new Focus, like the late 2000's focus and like the standard model without power windows.

You can go the used Mazda Miata route with something that is snappy and light like the Pixy or the OpenMV M7. They are cheap and they work really well but you gotta really get under the hood to properly use them and maintain them.

You could go the new model Chevy Camaro route and get something like the JeVois. It's going to take some aftermarket addons to get it rolling at full speed but once you get those bolted on and in place, it's going to do exactly what you need it to do and be fast about it.

You could also go the Ford Raptor route and get something like the Limelight. It does it all and then some and it's meant for this specific use case. It's expensive though.

You could go the Batmobile route and roll your own custom solution that helps you fight crime with something like a Jetson TX2 and a Stereolabs Zed camera. It's highly experimental and it might not work in all circumstances and you may have to abandon parts of it due to complexity and the need to use something more like a Batcycle.

Lastly, you could go with an M1 Abrams and find a sponsor who works with computer vision and get them to provide something like a special purpose industrial automation camera that works at 1000 frames per second and uses gigabit ethernet and comes with special proprietary software but it might be questionably legal as to your possession and usage of it.
__________________
"La mejor salsa del mundo es la hambre" - Miguel de Cervantes
"The future is unwritten" - Joe Strummer
"Simplify, then add lightness" - Colin Chapman
Clarke's Laws

Last edited by marshall : 12-19-2017 at 01:14 PM.
Reply With Quote
  #65   Spotlight this post!  
Unread 12-19-2017, 01:32 PM
gerthworm's Avatar
gerthworm gerthworm is offline
Making the 1's and 0's
FRC #1736 (Robot Casserole)
Team Role: Mentor
 
Join Date: Jan 2015
Rookie Year: 2015
Location: Peoria, IL
Posts: 706
gerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond repute
Re: JeVois Smart Machine Vision

Quote:
Originally Posted by marshall View Post
You could also go the Ford Raptor route and get something like the Limelight. It does it all and then some and it's meant for this specific use case.
Ford Raptors have a "specific use case"?
Reply With Quote
  #66   Spotlight this post!  
Unread 12-19-2017, 01:36 PM
marshall's Avatar
marshall marshall is offline
Online Stripe & Drape Sales
FRC #0900 (The Zebracorns)
Team Role: Mentor
 
Join Date: Jan 2012
Rookie Year: 2003
Location: North Carolina
Posts: 3,018
marshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond reputemarshall has a reputation beyond repute
Re: JeVois Smart Machine Vision

Quote:
Originally Posted by gerthworm View Post
Ford Raptors have a "specific use case"?
It can haul stuff and it hauls stuff. It's not meant for everyone.
__________________
"La mejor salsa del mundo es la hambre" - Miguel de Cervantes
"The future is unwritten" - Joe Strummer
"Simplify, then add lightness" - Colin Chapman
Clarke's Laws
Reply With Quote
  #67   Spotlight this post!  
Unread 12-19-2017, 01:37 PM
gerthworm's Avatar
gerthworm gerthworm is offline
Making the 1's and 0's
FRC #1736 (Robot Casserole)
Team Role: Mentor
 
Join Date: Jan 2015
Rookie Year: 2015
Location: Peoria, IL
Posts: 706
gerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond repute
Re: JeVois Smart Machine Vision

Quote:
Originally Posted by marshall View Post
...
https://www.youtube.com/watch?v=rKtFJxGzJX8
Reply With Quote
  #68   Spotlight this post!  
Unread 12-19-2017, 02:04 PM
asid61's Avatar
asid61 asid61 is offline
Design Simple
AKA: Anand Rajamani
FRC #1072 (Harker Robotics)
Team Role: Mentor
 
Join Date: Jan 2014
Rookie Year: 2013
Location: Cupertino, CA
Posts: 3,084
asid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond reputeasid61 has a reputation beyond repute
Re: JeVois Smart Machine Vision

Quote:
Originally Posted by Cog View Post
So as someone who has only tangential experience programming, can anyone sum up for me the differences between Pixy and JeVois?

Both seem pretty similar at first glance, being small compact cameras with on-board vision processing.
Pixy just detects blobs of colors and needs an Arduino to use well. It is sensitive to light level. JeVois can be made insensitive to light level and can communicate with the RoboRIO over USB or Serial, but needs more programming. It can perform much better than a Pixy fairly easily.
__________________
Team 1072 2017-present
Team 299 2017
Team 115 2013-2016 (student)

2018 Davis Finalists (w/ 6474 and 3880), 2018 Roebling Winners (w/ 3476, 1323, and 1778)

Reply With Quote
  #69   Spotlight this post!  
Unread 12-19-2017, 02:05 PM
Tom Bottiglieri Tom Bottiglieri is offline
Registered User
FRC #0254 (The Cheesy Poofs)
Team Role: Engineer
 
Join Date: Jan 2004
Rookie Year: 2003
Location: San Francisco, CA
Posts: 3,303
Tom Bottiglieri has a reputation beyond reputeTom Bottiglieri has a reputation beyond reputeTom Bottiglieri has a reputation beyond reputeTom Bottiglieri has a reputation beyond reputeTom Bottiglieri has a reputation beyond reputeTom Bottiglieri has a reputation beyond reputeTom Bottiglieri has a reputation beyond reputeTom Bottiglieri has a reputation beyond reputeTom Bottiglieri has a reputation beyond reputeTom Bottiglieri has a reputation beyond reputeTom Bottiglieri has a reputation beyond repute
Re: JeVois Smart Machine Vision

Quote:
Originally Posted by Cog View Post
So as someone who has only tangential experience programming, can anyone sum up for me the differences between Pixy and JeVois?

Both seem pretty similar at first glance, being small compact cameras with on-board vision processing.
The Pixy cam is for all intents and purposes a closed source device*. You can use their desktop app to configure what colors you want to track, filter color blobs based on size, and set up what the output should look like (SPI, analog signal, etc). The resolution of the image is small and you have limited options for what sort of shape based filtering you can do. This makes things like determining which thing is the target and which thing is the giant screen displaying the same color as the target hard to do reliably. On top of this, we have found the persistent settings on the Pixy camera to be unreliable, and have eliminated it from contention for use on our robots.

The Jevois camera is a system that is meant to be modified. You will write/find code that can track the target and upload it to the cameras onboard memory. Once the camera boots, it will run your custom code and output its results over a hardware or USB serial connection. You control all of the things the camera code is doing from setting camera hardware settings (exposure, AWB, etc) to finding targets to filtering targets to serializing results and sending them over the serial connection.

Our current strategy if 2018 requires vision tracking is to let our students develop a jevois camera module, and if that does not perform at the level it should (for whatever hardware reasons we have yet to find), go ahead and buy a Limelight camera.


*The code that runs on the pixy is indeed open to view, but the tracking code itself is part of a firmware image which would require you to build a system image and understand most of their software stack. I believe this is outside the scope of what an FRC should need to do.
Reply With Quote
  #70   Spotlight this post!  
Unread 12-19-2017, 03:14 PM
billbo911's Avatar
billbo911 billbo911 is offline
I prefer you give a perfect effort.
AKA: That's "Mr. Bill"
FRC #2073 (EagleForce)
Team Role: Mentor
 
Join Date: Mar 2005
Rookie Year: 2005
Location: Elk Grove, Ca.
Posts: 2,662
billbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond reputebillbo911 has a reputation beyond repute
Re: JeVois Smart Machine Vision

Quote:
Originally Posted by Tom Bottiglieri View Post
...Our current strategy if 2018 requires vision tracking is to let our students develop a jevois camera module, and if that does not perform at the level it should (for whatever hardware reasons we have yet to find), go ahead and buy a Limelight camera...
This is a perfect example of Poof logic. Simple, straight forward, easy to understand, robust.
__________________
CalGames 2009 Autonomous Champion Award
2012 Sacramento Semi-Finals, 2012 Sacramento Innovation in Control Award, 2012 SVR Judges Award.
2012 CalGames Autonomous Challenge Award winner ($$$).
2014 2X Rockwell Automation: Innovation in Control Award (CVR and SAC).
Curie Division Gracious Professionalism Award.
2014 Capital City Classic Winner AND Runner Up. Madtown Throwdown: Runner up.
2015 Innovation in Control Award, Sacramento.
2016 Chezy Champs Finalist, 2016 MTTD Finalist
2017 Utah Regional Winner!, Sacramento Finalist
Innovation in Control Newton/Carver Divisions , Newton #5 Captain
2018 WFFA Sacramento, Creativity Award Galileo-Robling Divisions
Reply With Quote
  #71   Spotlight this post!  
Unread 12-19-2017, 04:33 PM
AllenGregoryIV's Avatar
AllenGregoryIV AllenGregoryIV is offline
Engineering Coach
AKA: Allen "JAG" Gregory
FRC #3847 (Spectrum)
Team Role: Coach
 
Join Date: Jul 2008
Rookie Year: 2003
Location: Texas
Posts: 3,049
AllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond repute
Send a message via AIM to AllenGregoryIV
Re: JeVois Smart Machine Vision

Quote:
Originally Posted by gerthworm View Post
I won't claim this is the best way, but we've been having luck so far using the built-in WPILib functions, and have done some abstraction for running commands, checking for OK with timeout, and parsing for packets delimited by {}. Seems to be working, at least.

https://github.com/RobotCasserole173...Interface.java
Thanks for this, it does seem to work. Not sure why my original attempts which were very similar weren't working.
__________________

Team 647 | Cyber Wolf Corps | Alumni | 2003-2006 | Shoemaker HS
Team 2587 | DiscoBots | Mentor | 2008-2011 | Rice University / Houston Food Bank
Team 3847 | Spectrum | Coach | 2012-20... | St Agnes Academy
LRI | Alamo Regional | 2014-20...
"Competition has been shown to be useful up to a certain point and no further, but cooperation, which is the thing we must strive for today, begins where competition leaves off." - Franklin D. Roosevelt
Reply With Quote
  #72   Spotlight this post!  
Unread 12-19-2017, 05:33 PM
gerthworm's Avatar
gerthworm gerthworm is offline
Making the 1's and 0's
FRC #1736 (Robot Casserole)
Team Role: Mentor
 
Join Date: Jan 2015
Rookie Year: 2015
Location: Peoria, IL
Posts: 706
gerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond reputegerthworm has a reputation beyond repute
Re: JeVois Smart Machine Vision

Quote:
Originally Posted by AllenGregoryIV View Post
Thanks for this, it does seem to work. Not sure why my original attempts which were very similar weren't working.
Awesome! Glad to help!
Reply With Quote
  #73   Spotlight this post!  
Unread 12-19-2017, 08:18 PM
AllenGregoryIV's Avatar
AllenGregoryIV AllenGregoryIV is offline
Engineering Coach
AKA: Allen "JAG" Gregory
FRC #3847 (Spectrum)
Team Role: Coach
 
Join Date: Jul 2008
Rookie Year: 2003
Location: Texas
Posts: 3,049
AllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond reputeAllenGregoryIV has a reputation beyond repute
Send a message via AIM to AllenGregoryIV
Re: JeVois Smart Machine Vision

In Bill's guide he mentions sending out JSON from the JeVois and parsing it on the RoboRIO which I think makes the a lot of since. As we try to standardize this process for teams, is there a preferred JSON library for the RoboRIO? GSON or Jackosn?
__________________

Team 647 | Cyber Wolf Corps | Alumni | 2003-2006 | Shoemaker HS
Team 2587 | DiscoBots | Mentor | 2008-2011 | Rice University / Houston Food Bank
Team 3847 | Spectrum | Coach | 2012-20... | St Agnes Academy
LRI | Alamo Regional | 2014-20...
"Competition has been shown to be useful up to a certain point and no further, but cooperation, which is the thing we must strive for today, begins where competition leaves off." - Franklin D. Roosevelt
Reply With Quote
  #74   Spotlight this post!  
Unread 12-19-2017, 08:38 PM
noah.gleason's Avatar
noah.gleason noah.gleason is online now
Registered User
FRC #0449 (The Blair Robot Project)
Team Role: Programmer
 
Join Date: Mar 2017
Rookie Year: 2016
Location: Bethesda, MD
Posts: 139
noah.gleason has a reputation beyond reputenoah.gleason has a reputation beyond reputenoah.gleason has a reputation beyond reputenoah.gleason has a reputation beyond reputenoah.gleason has a reputation beyond reputenoah.gleason has a reputation beyond reputenoah.gleason has a reputation beyond reputenoah.gleason has a reputation beyond reputenoah.gleason has a reputation beyond reputenoah.gleason has a reputation beyond reputenoah.gleason has a reputation beyond repute
Re: JeVois Smart Machine Vision

Quote:
Originally Posted by AllenGregoryIV View Post
In Bill's guide he mentions sending out JSON from the JeVois and parsing it on the RoboRIO which I think makes the a lot of since. As we try to standardize this process for teams, is there a preferred JSON library for the RoboRIO? GSON or Jackosn?
We've based our robot code off of Jackson, but to use most of the Jackson annotations you need to have introspection stuff that's not in the compact JDK the Java Install tool uses by default. I don't know if this will be different when the Java installer gets deprecated in 2018, but if you want to use Jackson and test it now, I made a fork of the Java installer that allows you to install the full JDK.

EDIT: If it's possible to put libraries on the JeVois, I'd recommend using Google's Protocol Buffers instead of json, they're very good for this use case.

Last edited by noah.gleason : 12-19-2017 at 08:57 PM. Reason: Add links
Reply With Quote
  #75   Spotlight this post!  
Unread 12-19-2017, 09:26 PM
kylelanman's Avatar
kylelanman kylelanman is offline
Programming Mentor
AKA: Kyle
FRC #2481 (Roboteers)
Team Role: Mentor
 
Join Date: Feb 2008
Rookie Year: 2007
Location: Tremont Il
Posts: 248
kylelanman is a splendid one to beholdkylelanman is a splendid one to beholdkylelanman is a splendid one to beholdkylelanman is a splendid one to beholdkylelanman is a splendid one to beholdkylelanman is a splendid one to beholdkylelanman is a splendid one to beholdkylelanman is a splendid one to behold
Re: JeVois Smart Machine Vision

Quote:
Originally Posted by Tom Bottiglieri View Post
Our current strategy if 2018 requires vision tracking is to let our students develop a jevois camera module, and if that does not perform at the level it should (for whatever hardware reasons we have yet to find), go ahead and buy a Limelight camera.
Why the shift away from your smart phone solution? Given the initial time investment that has already been put into your android app and rio side code isn't tweaking the code to accommodate a new target straight forward and quick? The price point of a capable smart phones is significantly cheaper than the Limelight.

I'm mostly curious if you ran into significant problems with the Android solution that you are trying to combat with the Jevois / Limelight?
__________________







"May the coms be with you"

Is this a "programming error" or a "programmer error"?
Reply With Quote
Reply


Thread Tools
Display Modes Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -5. The time now is 07:58 PM.

The Chief Delphi Forums are sponsored by Innovation First International, Inc.


Powered by vBulletin®
Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.
Copyright © Chief Delphi