Wednesday, November 24, 2010

Field Geology the Hard Way

This past weekend I was in the Denver area to celebrate my neice's wedding.  Congratulations Kimberly and James!  We stayed near the Colorado School of Mines and had a few hours to walk around.  The Geology Museum was open and we spent an hour wandering around.  Highly recommended!

All kinds of cool exhibits on mining, radioactivity, fossils, UV fluorescent minerals, and gorgeous gemstones.  Walking downstairs you first encounter a pretty nifty exhibit on meteorites too.  Wandering a bit more I came across this.

Moon rocks are on exhibit in many museums but somehow I don't think I have ever seen one in person.  This is one of the Apollo 17 Goodwill pieces - two given to each state and some 135 countries.  I did some reading about these Goodwill rocks and was saddened to learn that of the 240 samples given to international governments some 180(!) are missing or unaccounted for.  Of the 100 given to the US states ... 33 are "unaccounted for".

Even this little piece has a story.  It had gone missing for decades but thanks to a lost moon rock sleuthing project run by Professor Joseph Gutheinz at the University of Phoenix it was "discovered" this summer (2010) in the possession of ex-Gov. John Vanderhoof, who left office in 1976.  The official press release covering its return to the people and display at the School of Mines states that it "had been in the safekeeping of former Gov. John Vanderhoof".  Fair enough - at least he gave it back.  And just in time for me - it was put on display only a couple months ago.

(For the record I think Joseph Gutheinz is a hero.  He assigns his students the task of hunting down these missing rocks and has been instrumental in the recover of many.)

The rock is encased in a sphere of acrylic mounted on a cheesy wooden plaque, behind a plexiglass window, inside a hefty safe.  Good idea as I so wanted to touch it.  And the safe also keeps ex-Governors from "helping".  I stood and stared trying to take in what I was seeing in that unassuming chunk of moon.  These rocks are of remarkable scientific value.  And they represent so much more.

There are three ways to get moon rocks:

1.)  Moon rocks can be found on earth, shards blasted from cratering events and finding their way to us as lunar meteorites.  I heard somewhere that the prevailing thought in the 1800's was that all meteorites were of lunar origin.  Today only a hundred some odd meteorites of lunar origin are known - the vast majority getting their start as meteoroids.  As the sites chosen for the Apollo missions have been shown to be geologically anomalous, these rocks from random impact sites on the moon have special scientific relevance. 

Certainly cheaper than going there

2.) And then there was the Soviet era robotic sample return mission(s) - which in itself is pretty amazing especially for the 1970's.  The three successful Lunas brought back a total 300 grams of samples. Here is a drawing of Luna 16, the first to bring back the goods for the Soviets.

1950's SciFi Cover Art?  Nope.  Luna sample return lander.

3.) Then there is the hard way.  Over 380 kg in all. 

Astronaut Charles Duke all dusty from a hard days field work.
I guess that is why I stared at that little piece of moon for a while.  Transfixed.

Some things really are worth doing.

I'm glad we did.

Sunday, November 14, 2010

Learning from the Sisters

As you can tell from my previous posts I've been spending a lot of time learning the basic fundamentals of PixInsight.  Using the Pleiades data that I gathered in the October new moon cycle, I've come to the conclusion that the Seven Sisters are an ideal training ground for astroimage processing.  Nice color, detailed nebulosity without being overly tricky (ala M42), some star color, and brightness variations to throw a little challenge into the mix.  A zillion amateur and professional images available on the net are also a help.  In other words a little bit of everything and nothing too difficult.

The fact that there is a detailed processing example on the PixInsight website helped a bit too.  Here is my current, um, version.  As always, click to flickr.

Encouraged by my early progress (see a couple posts back) I subsequently languished in a discouraging cycle of dead ends and restarts.  Trying more sophisticated tools against my image lead to worse results.  I finally discovered the NoiseEvaluation script and found that I was inexorably raising the noise floor as I processed deeper.  Even though I was keeping the data linear for as long as possible things just plain got noisier.  My tendency was to overstretch and otherwise use too aggressive settings and got to a place where even the sophisticated noise reduction tools in PI couldn't compensate.

Eventually I tackled learning how to do all the calibration processing in PI as well.  This is an area where PI does need some workflow streamlining - not for the faint of heart.  Nebulosity is much much easier.  But the tools in PI are powerful.  After I plowed through it I was able to get my subs to about 5-8x lower sigma SNR.  Pretty cool.  So now I was able to start post-processing with more headroom.

And well, I did only have an hour worth of exposure time.  I really shouldn't expect so much.  Two minute subs too.  But I think it is a very worthwhile learning experience to work with the same data set over and over to see just how much you can squeeze out of the noise.  Experiment often. Reprocess much. You will learn a lot which you can bring to bear on your subsequent images.

I have about 4 or 5 other image sets waiting to be processed but I've patiently spent hours and hours of time with the Seven Sisters while I learn PixInsight fundamentals.  Oh so pretty and alluring they are.  Not entirely painful company!


Taken at Lake San Antonio, CA  October 8, 2010
Canon T2i (stock), Astrotech AT80LE, 480mm  f5
Astrophysics AP900 mount

Guiding: SSAG/50mm Orion finder, PhD
30 120sec exposures @ ISO1600
94 darks
14 flats (1sec ISO 100

128 bias frames
calibration, registration and post-processed in PixInsight

Processing Partners

I love PixInsight.  OK, perhaps I'm getting a bit ahead of myself - after all it is those heady and exhilarating early days of relationship.  Let's just say I think this just may be the one.  We are going steady right now and I plan to pop the question even before the trial license period expires on 07:45:33 UTC December 12, 2010. 

This will not be an exclusive relationship.  Right tools for the job and all that.  I am on very good terms with my previous [processing] partner (Photoshop) and hope we will remain close friends.  We can spend quality time together on my terrestrial photography and I am sure that there may a few tricks that PS can do easier.  For those tasks we'll just go out for coffee.  But for astrophotography PS could never kindle my passion like PixInsight does.  In her heart of hearts I'm sure PS knew this day was coming. 

Contrary to popular belief, PixInsight is not difficult to get to know.  Exotic and perhaps a bit dangerous but that is part of the allure.  Algorithmic and rational.  Holistic and deep.  She doesn't paint herself or your images.  Yet watch her reveal the inner beauty lurking just above your noise threshold.  There is no user manual (yet) but this helps forge a deeper understanding right from the beginning.  And there are many who know her mysteries and are happy to share their advice.

If you are looking for a superficial one button relationship - don't bother.  You will have to invest some quality time but oh so well worth it.  You will get to learn strange and foreign sounding terms like ATrousWavelets and HDRWTransform.  You will get frustrated. Yes you will even fight at times.  But oh so richly rewarding.  If you want to inject a some fun and excitement into your processing experience ... on the next full moon invite PixInsight out for a night on the town.  You won't regret it.

Sunday, November 7, 2010

New Insight

It's been a while posting but that doesn't mean I've neglected wearing the hair shirt.

I went to the 2010 Advanced Imaging Conference which was held over the weekend of Oct. 22.  I wavered about attending.  On the positive side, it was local - and it had a really great line up of speakers. On the down side it was rather spendy and since my regular imaging pal wasn't going I wasn't looking forward to being the token "Novice Hair Shirt".

In the end it was a great experience.  Met some really nice people and the talks were entertaining and informative.  I'd recommend the conference to anyone interested in astro imaging no matter what your experience level.  Oh, and they have a nice vendor fair which was well attended - all kinds of cool toys to look over.

Rogelio Andreo (see his Deep Sky Colors web site at right sidebar) gave a fabulous talk on his imaging techniques as well as some food for thought on image composition and goals.  It was the latter that made the biggest impression on me but I'll save those thoughts for another post.   On the technical side he talked about processing images with PixInsight in which he described some of the workflow and concepts.  As he said over and over again ... "not better, just different".

A week ago I finally downloaded the trial license of PixInsight and have been slogging my way up the learning curve.  I had been working in photoshop on an image of the Pleiades I took at Calstar, so chose that as my first go with PixInsight.  Below is a comparison of the same data - preprocessed in Nebulosity and the post-processed with PS and PI.

I know if I had worked a bit longer on the PS version I would have done significantly better than the above attempt - which admittedly kind of stinks.  And the PI version is stretched till the pixels screamed and bled noise all over the image.  But I could, in a rather short time - maybe a couple hours working through some of the tutorials -  come up with a result in PI that bettered my PS attempt.

Being relatively new to image processing I haven't yet locked down my synaptic pathways in the ways of PS.  This is one of those rare situations when not knowing much may be an advantage.  I can learn a new way of accomplishing something without having the disadvantage of going through a translation step.

 So I've put aside PS for now - and decided to spend some time learning a completely new processing methodology - before becoming proficient at the one I've put months of learning into.  Beautiful.  How hair shirt is that?!