[robocup-small] Unique patterns - SSL Vision

[Michael Bleier]

Hi Miguel and Joydeep,

We also had troubles with this kind of setup a couple times. The
parameters of SSL-Vision are optimized for around 4m camera height. The
software has thresholds that rejects Blobs that are too big in the
camera frame.

Here are a few ideas, which might cause the problem:

(1) There are a couple of threshold for the maximum size of the color
blobs and also the distance between the different blobs. You can find
them in the settings tree

Global > Team Config > Teams > Standard2010 >
Pattern Fitting or Center/Other Marker Settings

(2) There is a magic constant in the code for the distance in the search
tree for the green/pink color patches. Maybe this distance is too
restrictive.

file: src/shared/cmpattern/cmpattern_teamdetector.cpp
code line: reg_tree.startQuery(*reg,20.0);

Best regards,
Michael

On 10/08/2013 05:12 PM, Joydeep Biswas wrote:
> Hi Miguel,
> 
> Good to hear about your entrance to the SSL! :)
> 
> To help debug this issue, please send me two images captured by your
> cameras, one with no robots on the field (so that I can use it for
> geometry calibration), and the second with some robots on the field.
> For instructions on how to capture these images, see the section named
> "Recording Images to Files" on the wiki:
> https://code.google.com/p/ssl-vision/wiki/CapturingImages
> 
> Regards,
> Joydeep
> 
> On Sun, Oct 6, 2013 at 4:01 PM,  <miguel.solis at usm.cl> wrote:
>> Hi everyone,
>>
>> i'm from a new team excited for finishing our current work and looking to
>> participate in the next robocup competition. For different reasons, we have
>> a very very reduced space for making tests (the current area of the field we
>> have is of 1.6 mts x 3 mts) and the camera is at a height of 2 mts, and we
>> have some issues with ssl-vision and we'd like to know your opinions about
>> what possibly is going wrong...
>>
>> we're able to identify blue and yellow team (verifying it on the graphical
>> client), but when we check the box of "unique patterns" the robots suddenly
>> disappear and we aren't able to identify the id of each robot.
>>
>> Is it possible to fail because the field test is too small ?, because the
>> distortion set at calibration ? (distortion = 0.25) ...
>>
>> we are working with the standard pattern, the pattern image is ok and its
>> path is correctly defined (no error at loading the pattern image is
>> shown)... we'd appreciate any help or advise,
>>
>> Thank you,
>>
>> Best regards,
>>
>> Miguel Solis
>> IRSS Deluxe Team
>> _______________________________________________
>> robocup-small mailing list
>> robocup-small at cc.gatech.edu
>> https://mailman.cc.gatech.edu/mailman/listinfo/robocup-small
> _______________________________________________
> robocup-small mailing list
> robocup-small at cc.gatech.edu
> https://mailman.cc.gatech.edu/mailman/listinfo/robocup-small
> 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 901 bytes
Desc: OpenPGP digital signature
URL: <https://mailman.cc.gatech.edu/pipermail/robocup-small/attachments/20131008/f3e1019a/attachment.bin>