Cloud sensor

Technical Support for the BRT telescope and website
rjmodic
Posts: 88
Joined: Sun Dec 23, 2012 11:09 pm

Cloud sensor

Postby rjmodic » Tue Aug 18, 2015 4:23 am

The last 2 ro 3 nights it seems that the cloud detection sensor is getting a lot of false positives. Skies have been mostly clear and, unfortunately, a lot of telescope time has been lost. Someone should look into this.

Thanks,

Bob

rjmodic
Posts: 88
Joined: Sun Dec 23, 2012 11:09 pm

Re: Cloud sensor

Postby rjmodic » Fri Aug 21, 2015 10:57 pm

One other thing I've noticed about the cloud sensor is that, if skies are cloudy at shutdown time, then the cloudy value remains stuck throughout the next day and well after sunset the next night. It doesn't reset itself until the sky gets dark enough for the pole star camera to see stars. This results in the scope getting a late start by about an hour. The cloudy value should reset at the end of each night instead.

Bob

rjmodic
Posts: 88
Joined: Sun Dec 23, 2012 11:09 pm

Re: Cloud sensor

Postby rjmodic » Tue Sep 01, 2015 6:09 pm

There were more false positives from the cloud detection sensor last night. Skies were clear each time. Is there a problem with the pole star camera occasionally returning bad frames?

Bob

User avatar
Ed
Site Admin
Posts: 320
Joined: Wed Sep 21, 2011 2:18 pm

Re: Cloud sensor

Postby Ed » Tue Sep 01, 2015 8:21 pm

It's just a surprisingly difficult hard problem to solve with a computer. People usually get PhD's for solving problems like this one (after dedicating three years or so of their lives to it).

Humans are so very good at taking in visual information and effortlessly processing it that we take it for granted. Image recognition is as about a hard a computer science problem as exists.

False positives on cloud detection is better than false negatives I think (which was the problem previously). If the telescope doesn't observe for half an hour when it could, that is better than observing for half an hour when it shouldn't. It wastes the same amount of time at the end of the day (because half an hours work extra needs to be done another night), but the latter wastes a load of our users' time while they wait again for their picture to be taken, and is frustrating.

Anyway, I'm working on it (as much as my limited expertise allows). It is slowly getting better.

rjmodic
Posts: 88
Joined: Sun Dec 23, 2012 11:09 pm

Re: Cloud sensor

Postby rjmodic » Tue Sep 01, 2015 11:00 pm

Ed,

Thanks for the reply. As I type this (21:50 UT), another false positive occurred twice in the last half hour. A 180 sec BVR job failed twice as a result of this.

How often does the cloud sensor/camera take a measurement? Will a single bad measurement be enough to trigger an error conditon or does the system need multiple detections in a row? Maybe a single bad reading should be ignored (as long as its not raining).

This problem seems to have developed over the last month.

Bob

David Conner
Posts: 133
Joined: Wed May 26, 2010 11:26 pm

Re: Cloud sensor

Postby David Conner » Wed Sep 02, 2015 12:20 am

Hi,

There often seems to be a short 'blip' on the cloudyness graph round about 22.30 / 23.00.

David.

David Conner
Posts: 133
Joined: Wed May 26, 2010 11:26 pm

Re: Cloud sensor

Postby David Conner » Thu Sep 10, 2015 9:46 pm

Any progress on the false positives on the cloud sensor? They are predictable and occur on almost every night.

David.

rjmodic
Posts: 88
Joined: Sun Dec 23, 2012 11:09 pm

Re: Cloud sensor

Postby rjmodic » Mon Sep 14, 2015 10:21 pm

Just to echo what David said, there are false positives almost every night with the cloud sensor. Since many of these are happening during the first hour of operation, many early evening jobs are not getting done.

The Dec. homing issue has also returned, also causing delays during the first hour of many nights.

Please do something about these problems.

Thanks,

Bob

rjmodic
Posts: 88
Joined: Sun Dec 23, 2012 11:09 pm

Re: Cloud sensor

Postby rjmodic » Mon Sep 14, 2015 10:32 pm

In the last hour (~21h UT, 9/14), there have now been 4 false positives with the cloud sensor. The Dec. is not homed properly and the dome is blocking half the FOV of the ConstellationCam!

Please do something about this ASAP!

Thanks,

Bob

User avatar
Chris
Site Admin
Posts: 147
Joined: Tue Jan 15, 2008 4:17 pm
Location: BRT Headquarters
Contact:

Re: Cloud sensor

Postby Chris » Fri Oct 02, 2015 12:50 pm

I investigated the webcam software which supplies images to the cloud detection software - there doesn't seem to be anything wrong with the camera switching logic. So, I have added even more delay to the video switcher to allow it to switch cameras before obtaining a frame. A while back we moved from a switch box which replies when it has switched, to one that doesn't - which is why the post-switch delay time is guesswork. It looks to have had a positive effect but if you see it doing the same again, let me know.


Return to “Technical Support”

Who is online

Users browsing this forum: No registered users and 1 guest

cron