Testing OpenCV filters w/ Custos Mundus' Brain

This is the IR pattern projected by the kinect as viewed from the NIR webcam on Custos Mundus. 

Pretty cool pattern and an unforeseen touch that made my day when I saw it while testing out filters in OpenCV.

 

 

 

 

 

 

Holy Saturation Value, Batman! That's right,  HSV is back and spreading like an epidemic. Its pretty great really. picks up some great depth and shades. Its also closely related to the false color NDVI readings that we want to take with Custos and the rest of his plantoid cronies.

 

 

 

 

 

Detector currently bring the OpenCV service down quite a bit, but is really active. 

 

 

 

 

 

 

not quite as active as the Fauvist filter though.

 

Which is really really fun to watch. Picks up lots of edge detail and movement. 

 

 

 

 

 

 

 

 

Any of these filters would be awesome to have projected back out onto the range of vision. This is helped a bit by the nifty effect left by the kinect IR pattern, but I added a mini projector for a bit of display fun for the robot to have. Maybe it can shine the pattern effect onto its field of view and show its own readings to bystanders or others "in the room". 

Thanks all here and especially GroG, everyday is a new awesome thing.


Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
GroG's picture

My God—it's full of stars!

I hyper-linked each of your fine screen captures because they are just very wicked cool.  Great post.  It got me thinking too...  Do you know the conceptial difference between the HSV filter and the false coloring?  I suspect in order to find healthy plant growth the NIR python filters is looking for specific spectrum yes?  I am probably over simplifying ?   And thanks for sharing and being part of the collective .. HAIL BEPSL !

 

DJUltis's picture

HSV with Infragram

The interesting part concerning getting the right spectrums is that the deep blue filter that I hacked the webcams with gives us the right spectrums to analyze. with the HSV filter, we can see it working, though I'll put an unhacked webcam on there so that we can see a carrier aswell.  The InRange filter also worked well for filtering spectrum for robotic purposes as well. 

Here's my morning reading:

http://plots2.media.mit.edu/notes/warren/08-24-2013/infragrammar-with-hsv-color-model

Found this thismorning as well, basically same thing we can do with live MRL viewer: http://plots2.media.mit.edu/notes/warren/08-24-2013/infragrammar-with-hsv-color-model

This GIMP video guide expresses false colr techniques to process images, and it really does seem like the HSV maybe just wants to be tuned to specific settings. 

 

 

DJUltis's picture

More on HSV

This is a good thread as well. 

https://groups.google.com/forum/#!topic/plots-infrared/3POpoFGOxV4

borsaci06's picture

great post...

liked the post and your efforts in NIR imaging.... have you found the Rasco2007 blue filter?... I couldnt find it anywhere... try imaging some living plants just to see the effects with opencv filters... there is a Plugin (  http://publiclab.org/wiki/photo-monitoring-plugin ) for Fiji software to add the IR image instead of Red in RGB to prepare a composite NDVI... and also as you might already know, PublicLabs has an online tool for that which can be accessed via their page..  I dont have that software, I prefer working with Photoshop... if Grog can do something with OpenCV filters and MRL access to these online tools or I can try the same approach in Photoshop by writing some macros... great post David..

 

DJUltis's picture

Yeah man,

I've got a few sheets. Need some? Always willing to help the fellows out...

borsaci06's picture

rosco 2007 blue

Thx... would like to have one... all I need is a small sheet... but how shall u send it to me?... one small piece to cover the lens and a bigger one to use as custom white balance background...