Mt. Diablo change monitoring media coverage, alignment software

[Ken-ichi Ueda]

Hey folks,

Just thought I'd let you know that Nerds for Nature's implementation of this idea has been getting some media coverage:

Grist: http://grist.org/business-technology/oh-snapchat-your-smartphone-just-became-a-climate-scientist/

Local CBS: http://sanfrancisco.cbslocal.com/2014/05/09/nature-nerds-photograph-monitor-mount-diablo-morgan-fire/

KQED: http://blogs.kqed.org/science/2014/04/24/hikers-use-smartphones-to-capture-fire-recovery-on-mt-diablo/

Also, Ken McGary of N4N recently pitched this at EcohackSF and collaborated with programmer Petr Glotov to create https://github.com/pglotov/TimeLapser, a program that automatically aligns photos without the need to preselect control points. It's based on a modified version of OpenCV, but there's a pull request pending: https://github.com/Itseez/opencv/pull/2720. It doesn't do any kind of spectral normalization or actually make a movie, but it's a significant step in the right direction.

You can see some of the photos from this project at http://bit.ly/morganfire. Still some work to do wrangling the data (that's on me, probably), but things seem to be going well! Once again, thanks to Sam Droege et al. for the original idea.

-ken-ichi

[Droege, Sam]

Ken-ichi....

Excellent work here.  

We will definitely play around with the TimeLapser your group put together and see if we can make it yelp.  

We also did some work this past year looking at how accurately PtGui corrected pictures across camera types.  Still have to tweak those data, but the bottom line was that there are likely going to be instances where certain cameras won't sync (at least with this sort of software) and that measurement error among those that do appears to be super tiny if you are using the same camera type, but there are "significant" differences among camera types.  However, because measurement error was so small within a particular camera almost any among camera differences become significant...thus we will have to play around a bit more to see how biologically significant such changes might be, could actually be insignificantly small biologically.

All good.

sam

Sam Droege  sdr...@usgs.gov                      
w 301-497-5840 h 301-390-7759 fax 301-497-5624
USGS Patuxent Wildlife Research Center
BARC-EAST, BLDG 308, RM 124 10300 Balt. Ave., Beltsville, MD  20705
Http://www.pwrc.usgs.gov

DAGAZ

Down the cathedral aisles
of deep and ancient forests
  bursts the morning fire.
Take heart, my Beloved.
What is Divine
   cannot be separated
from its Source
                  Ralph H. Blum

--
An open source community for collaborative development of change monitoring systems, tools and applications. People can do this right now using existing materials at single sites or they can organize networks of camera stations at scales of parks, cities, watersheds, counties, states, countries, or the world.
 
This is a presentation of an idea. Anyone can modify this in any way they like and implement it at any scale. No copyrights. No permissions needed. Just Do It.
---
You received this message because you are subscribed to the Google Groups "Monitor Change" group.
To unsubscribe from this group and stop receiving emails from it, send an email to monitor-chang...@googlegroups.com.
To post to this group, send email to monitor...@googlegroups.com.
Visit this group at http://groups.google.com/group/monitor-change.
For more options, visit https://groups.google.com/d/optout.

--

Bees are Not Optional

Apes sunt et non liberum

[Ken-ichi]

Hey Sam,

Cool. If you can get those results online (especially your methods)
when they're done that would be great. We're mostly interested in the
qualitative aspects of this project, but it would be really
interesting to know exactly how you're using PtGUI to make
quantitative comparisons between images. If we go down that path we'd
probably be more interested in pursuing open source solutions that
either use the underlying PT libs directly (e.g. Hugin) or make use of
other libraries like OpenCV (as TimeLapser does).

-ken-ichi