Page 1 of 1

Use of OpenCV

Posted: Sat Nov 18, 2017 8:29 pm
by kramkroc
Hi foks,

two of our teams were accepted for phase two and both are Life on Earth based projects revolving around capturing images on Pi NoIR and then crunching the images. I wondered whether use of OpenCV is permitted or if there are python resources people are aware of for edge detection, bright sport detection etc?

/Mark

Re: Use of OpenCV

Posted: Mon Nov 20, 2017 9:16 am
by Davespice
Hi there

OpenCV was used on the Astro Pi units during the Peake mission in 2016. So there's no problem with requiring it, we're expecting a lot of teams to want to use it.

Many thanks

Dave

Re: Use of OpenCV

Posted: Mon Nov 20, 2017 3:39 pm
by kramkroc
Hi Dave,

thanks again for that response. Just a quick question on versions. Will the latest be used (I think it is 3.3.0) or some other version? Had the joy of installing this the other day on my own pi so want to make sure I install the correct version on the Astro PI we got from you and the ESA.

/Mark

Re: Use of OpenCV

Posted: Mon Nov 20, 2017 3:40 pm
by kramkroc
arghhh...forgot to ask, but are there any samples of the images that are captured by the camera from the ISS? Want to calibrate our detection with some real examples.

/Mark

Re: Use of OpenCV

Posted: Tue Nov 21, 2017 8:58 am
by Davespice
Hi Mark

OpenCV is currently not installed on the Astro Pi units on orbit, so we will just install whatever is the latest version.

As a side note about the camera, please take note of this blog post. Particularly the blue filter (there will be one supplied with your Pi NoIR camera). The Blue filter will be installed on the Earth Observing Astro Pi unit for the duration of the mission. It basically stops visible red and green, but lets visible blue and infrared pass through into the camera. It permits a type of EO analysis of vegetation on ground. Infrared then goes into the red RGB channel of the image data.

We don't have any images taken from the ISS with this blue filter, but we do have some taken by that camera without any filter.
These are available here.

Note that:
  • these are low res images (640x480) but you'll be able to get much higher res images when you run your code (2592x1944).
  • the porthole is visible in the view, you may need to account for this and only examine the middle of the image
  • the porthole may not be in the same place, when your code runs - so don't hard code this - detect it and accommodate
If you need some sample images with the filter, they can be found on this flickr album.

Many thanks

Dave

Re: Use of OpenCV

Posted: Fri Nov 24, 2017 2:12 pm
by kramkroc
Wow. Some of those night time images are a challenge. In fact I've not found one yet that makes sense to the human eye :D

This is an example of what I mean:

https://github.com/astro-pi/enviro-pi/b ... 143009.jpg

The light from inside the ISS is reflecting off the glass and the underside of the pi is then all that is visible on the picture. Will the blue filter solve that problem?

/Mark

Re: Use of OpenCV

Posted: Sat Nov 25, 2017 8:40 pm
by Davespice
Hi Mark

Yes we were getting quite bad window reflections during night passes. This wasn't a problem for that experiment because it was programmed to discard night images anyway, but this logic was based on an average colour check and because of the reflection it meant that the night side images were kept.

We're currently working out the best way to ensure we don't get reflections like this when we deploy the student code in 2018. But for now you can work assuming you won't have to deal with reflections in your experiment.

Many thanks

Dave