Changing Active Window for HaViMo2?

Created by Prof. Hamid Moballegh of The Robocup Humanoid Team of Freie Universität Berlin - fumanoids.de
2 postsPage 1 of 1
2 postsPage 1 of 1

Changing Active Window for HaViMo2?

Post by Bae » Mon Dec 12, 2011 6:33 pm

Post by Bae
Mon Dec 12, 2011 6:33 pm

Hello Hamid,

In the documentation of the vision chip HV7131RP, the registers table on page 14 seemed to suggest that it may be possible to change the size and location of the Active Window (so that a full frame does not have to be used all the time). Do you think that we can do this via HaViMoGUI? (which is a great tool, thank you for sharing that).

The reason for my question is that I am interested in detecting "multiple" objects of the "same" color by using a smaller Active Window in conjunction with Pan-Tilt servos. And if I understand correctly from your HaViMo2 manual of March 10, 2010, it looks like that we can have statistics for only one object in each color? I would need statistics on several blobs of the "same" color (from the full frame), thus my need for smaller Active Windows and be able to vary its size/location.

I would appreciate any correction in my understanding or any suggestion/approach for this feature.

Regards,
Bae
Hello Hamid,

In the documentation of the vision chip HV7131RP, the registers table on page 14 seemed to suggest that it may be possible to change the size and location of the Active Window (so that a full frame does not have to be used all the time). Do you think that we can do this via HaViMoGUI? (which is a great tool, thank you for sharing that).

The reason for my question is that I am interested in detecting "multiple" objects of the "same" color by using a smaller Active Window in conjunction with Pan-Tilt servos. And if I understand correctly from your HaViMo2 manual of March 10, 2010, it looks like that we can have statistics for only one object in each color? I would need statistics on several blobs of the "same" color (from the full frame), thus my need for smaller Active Windows and be able to vary its size/location.

I would appreciate any correction in my understanding or any suggestion/approach for this feature.

Regards,
Bae
Bae offline
Newbie
Newbie
Posts: 4
Joined: Mon Dec 12, 2011 6:05 pm

Post by hamid_m » Mon Dec 19, 2011 4:39 pm

Post by hamid_m
Mon Dec 19, 2011 4:39 pm

Hi Bae,

The vision algorithm only works with a 160x120 YCrCb frame with a 1/6 sub-sampling. Any changes to the formatting registers may have unexpected results.

Best regards,
Hamid.
Hi Bae,

The vision algorithm only works with a 160x120 YCrCb frame with a 1/6 sub-sampling. Any changes to the formatting registers may have unexpected results.

Best regards,
Hamid.
hamid_m offline
Savvy Roboteer
Savvy Roboteer
Posts: 133
Joined: Thu May 03, 2007 4:56 pm


2 postsPage 1 of 1
2 postsPage 1 of 1