It is currently Thu, 23-11-17, 14:59 GMT

All times are UTC




Post new topic Reply to topic  [ 21 posts ]  Go to page 1, 2  Next
Author Message
 Post subject: EarthNight
PostPosted: Fri, 28-12-12, 18:35 GMT 
Offline
User avatar

Joined: Tue, 04-09-07, 21:55 GMT
Posts: 773
Location: N 42.38846 W 83.45456
Time to move the 2012 earthnight out of "Texture Tools" I am not using the TX tools so.....

using the 2006 Cal. ( f16) as a burn mask
and after adding the 2000 "gas flares" and "fire"

and setting NY NY to about 160 ( 140 to 190 over downtown)
--- the 2006 cal had NYNY at a 76 px value ---

i get this
Image

and Celestia screen shots
Image Image

I have a bit of a habit of just scrapping things and starting over
-- and this is a start over

removed the "earth map" from the background of the 2012 EO EarthNight from each of the RED,GREEN,BLUE layers
( the EO 2012 is normalized so
so subtract the value of the ice RED 15, GREEN 31, BLUE 82 and renormalize each layer

This moved the image into the 32 bit FLOAT
so doing a convert to UCHAR put the ( now negative earth map to ZERO)

burned each layer using a slightly blurred 2006 Cal.

Multiply one image by the other and divide by the max px value on the mask , this also moved the image back into FLOAT
Then adjusted exposure and gama to "0.4" ( for looks )

then back to 0 to 255 normalized UCHAR as the orig. image was .

Combined the three red,green,blue layers into a single image

the location of the MAX px value is still the same in this rework of the 2012 as was in the 2006 cal.

there are some artifacts that need removing
lights not in the 2006 cal map are a bit BRIGHT

--the cat is being a pest and is walking on the keybord
-- back later

_________________
"I don't pitch Linux to my friends, I let Microsoft do that for me."
Using OpenSUSE 42.1 & Scientific Linux 6.7


Last edited by John Van Vliet on Fri, 28-12-12, 22:54 GMT, edited 1 time in total.

Top
 Profile  
 
 Post subject: Re: EarthNight
PostPosted: Fri, 28-12-12, 19:49 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4510
Location: Hamburg, Germany
John Van Vliet wrote:
Time to move the 2012 earthnight out of "Texture Tools" I am not using the TX tools so.....


;-)

Quote:
using the 2006 Cal. ( f16) as a burn mask
and after adding the 2000 "gas flares" and "fire"

and setting NY NY to about 160 ( 140 to 190 over downtown)
--- the 2006 cal had NYNY at a 76 px value ---

i get this
Image

and Celestia screen shots
Image Image

I have a bit of a habit of just scrapping things and starting over
-- and this is a start over

removed the "earth map" from the background of the 2012 EO EarthNight from each of the RED,GREEN,BLUE layers
( the EO 2012 is normalized so
so subtract the value of the ice RED 15, GREEN 31, BLUE 82 and renormalize each layer

burned each layer using a slightly blurred 2006 Cal.
( adjusted exposure and gama for looks )
Multiply one image by the other and divide by the max px value on the mask

moved the gama to 0.4 ( looks )


the location of the MAX px value is still the same in this rework of the 2012 as was in the 2006 cal.

there are some artifacts that need removing
lights not in the 2006 cal map are a bit BRIGHT

--the cat is being a pest and is walking on the keybord
-- back later


Looking very good indeed.

F.


Top
 Profile  
 
 Post subject:
PostPosted: Fri, 28-12-12, 20:48 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4510
Location: Hamburg, Germany
OK,

I had a look at my 32k VT Earthnight map that I made myself a number of years ago.
I used as raw input the 32k radiance calibrated BW nightlights map and colored it with GIMP, using a truecolor reference photo taken by astronauts.

Image

During the GIMP coloration process I requested the luminosity to remain CONSTANT. This procedure avoids any color-wallpaper-type enhancements for which NASA imaging wizzards are so well known ;-)

Here is how the NY region looks like in my case:
Image

I noted this dark bluish background in your EarthNight images which I have avoided throughout. Here is a full-Earth shot:
Image

Anyway, you new maps look beautiful. For curiosity, I would still like to see the new EarthNight images with NYNY reduced to 76 px (instead of your present 160).

Fridger


Top
 Profile  
 
 Post subject: re
PostPosted: Fri, 28-12-12, 22:07 GMT 
Offline
User avatar

Joined: Tue, 04-09-07, 21:55 GMT
Posts: 773
Location: N 42.38846 W 83.45456
i just reduced the 64k to 32k and converted that vips format image to ppm and zipped it
-- 13,1 meg zip of a 1.5 GIG image
http://www2.zshare.ma/echenw9du6hj

this is BEFORE i added the gas flares and fires from the 2000 data and BEFORE i added 20% of the topo to the blue layer

NYNY is now at about 80
60 to 85 and a downtown Manhattan at 190( a spot )

this is NOT a strait burn from the 2012 using the 2006 mask , but a gama move

but it should be rather close to that seeing as i moved the gama to begin with .

I can recreate the direct burn of the 2012 map layers
without ANY corrections ( other then removing the 2012 earth map background)

if you want
BUT
that was very dark ( opinion - to dark )

i will have time late tonight

--- added later ---
Quote:
I noted this dark bluish background in your EarthNight images which I have avoided throughout. Here is a full-Earth shot:

yes there is some in it
here is a screen capture of the map with a X 6 px value exaggeration
( easy in Nip2 - move a slider )
Image

The Antarctic has a PX value of 21 to 23 in only the blue channel

_________________
"I don't pitch Linux to my friends, I let Microsoft do that for me."
Using OpenSUSE 42.1 & Scientific Linux 6.7


Top
 Profile  
 
 Post subject: RE
PostPosted: Sat, 05-01-13, 2:49 GMT 
Offline
User avatar

Joined: Tue, 04-09-07, 21:55 GMT
Posts: 773
Location: N 42.38846 W 83.45456
It will be a bit before the map is ready
I just got a mail from the Earth Observatory teem
they are putting together 16 bit data that is not so "stretched"

from that the 32 bit FLOAT 2006 Radiance Cal. can be matched better


I should have more info in about a week or so.

To quote
Quote:
I need some time to document the 16-bit data. Please ping me again if I don't send you a notice that it's posted by mid next-week.

The existing maps aren't spatially normalized, but they are highly stretched to show detail at both high and low light levels.

_________________
"I don't pitch Linux to my friends, I let Microsoft do that for me."
Using OpenSUSE 42.1 & Scientific Linux 6.7


Top
 Profile  
 
 Post subject: Re: RE
PostPosted: Sat, 05-01-13, 9:12 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4510
Location: Hamburg, Germany
John Van Vliet wrote:
It will be a bit before the map is ready
I just got a mail from the Earth Observatory teem
they are putting together 16 bit data that is not so "stretched"

from that the 32 bit FLOAT 2006 Radiance Cal. can be matched better


I should have more info in about a week or so.

To quote
Quote:
I need some time to document the 16-bit data. Please ping me again if I don't send you a notice that it's posted by mid next-week.

The existing maps aren't spatially normalized, but they are highly stretched to show detail at both high and low light levels.


That's interesting! Thanks for letting us know, John.

Looking forward to the results,
Fridger


Top
 Profile  
 
 Post subject: re
PostPosted: Wed, 16-01-13, 10:12 GMT 
Offline
User avatar

Joined: Tue, 04-09-07, 21:55 GMT
Posts: 773
Location: N 42.38846 W 83.45456
A screenshot of Detroit
using the 32 bit float earth night map
Image

it is an 86400 x 33600

same lat / long as the 2006 radiance calibrated
on the DMSP site

there is some noise and some cloud cover in the orig.
it needs a lot of work

_________________
"I don't pitch Linux to my friends, I let Microsoft do that for me."
Using OpenSUSE 42.1 & Scientific Linux 6.7


Top
 Profile  
 
 Post subject: Re: re
PostPosted: Wed, 16-01-13, 18:39 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4510
Location: Hamburg, Germany
John Van Vliet wrote:
A screenshot of Detroit
using the 32 bit float earth night map
Image

it is an 86400 x 33600

same lat / long as the 2006 radiance calibrated
on the DMSP site

there is some noise and some cloud cover in the orig.
it needs a lot of work


One can "smell" the amazing resolution. But for now things don't look brilliant yet on my brand new 24" super-performance monitor ;-) So agreed, quite a bit of work ahead here.

Thanks for sharing,

Fridger


Top
 Profile  
 
 Post subject:
PostPosted: Sat, 11-05-13, 2:40 GMT 
Offline
User avatar

Joined: Tue, 04-09-07, 21:55 GMT
Posts: 773
Location: N 42.38846 W 83.45456
just an update
i am making progress , there is A LOT of ground noise that needs removing .What is left is the Canadian north and some artifacts from that removal .

but screen shots of the Gama exaggerated B&W working on image ( 86400x43200 32 bit , value 0.0 to 1.0 )

Image Image Image


Top
 Profile  
 
 Post subject:
PostPosted: Sat, 11-05-13, 21:00 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4510
Location: Hamburg, Germany
Looks impressive! Did you clean with combinations of high and low passes (ISIS)?

f.

_________________
Image


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 13-05-13, 0:39 GMT 
Offline
User avatar

Joined: Tue, 04-09-07, 21:55 GMT
Posts: 773
Location: N 42.38846 W 83.45456
there are two types of noise
a "salt & pepper" ( easy to remove )
-- from memory
the city lights are from about 1 to 6000 px value
the s&p is from about -2000 to 16000 px value
then do to the sensitivity the moon light and normal sky glow lit up the ground

that needed to be removed with a "highpass" of the Earth texture ( red band )
subtracting it from the earthlights map

it is not 100% in north/ northwest Canada , that lighting looks to be aurora

the S&P noise is easy to remove with gmic ,in sections ( CImg.h has a 4 gig max , 32 bit coding )
the "-remove_hotpixels" is very much like the highpass/lowpass
BUT the isis tool uses a square boxcar that had to be large enough ( 13x13) that was leaving NASTY artifacts
this is using a heat flow pde to remove the salt and pepper
Code:
gmic -v 5 -type float 1.raw,86400,5400 -remove_hotpixels 5,0.75% -type float -o 1a.tiff

as seen in the orig*.cub
Image
then with the water masked and the land subtracted . It leaves the ambient lighting
-- Canada -
Image

that has to be removed by hand


Top
 Profile  
 
 Post subject:
PostPosted: Mon, 13-05-13, 8:25 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4510
Location: Hamburg, Germany
But it seems to me that you also removed the shine of the street lights? I usually watch the looks carefully when I am in a plane at night...

Fridger

_________________
Image


Top
 Profile  
 
 Post subject:
PostPosted: Thu, 16-05-13, 4:09 GMT 
Offline
User avatar

Joined: Tue, 04-09-07, 21:55 GMT
Posts: 773
Location: N 42.38846 W 83.45456
the "shine" ?
well the close ups above are from a the gamma moved orig. I moved to 0.25 so i can work on them in cinepaint .

_________________
"I don't pitch Linux to my friends, I let Microsoft do that for me."
Using OpenSUSE 42.1 & Scientific Linux 6.7


Top
 Profile  
 
 Post subject:
PostPosted: Thu, 16-05-13, 10:26 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4510
Location: Hamburg, Germany
What origin? sorry...

What I mean is that each streat light or other light shines. This means that it produces an area of decreasing light intensity around the proper light source. If one looks from an airplane (or satelite) onto a city, this coherent shine is very conspicuous, since there are many streetlights.
Too much noise filtering can easily also filter away this crucial diffuse shine.

Fridger

_________________
Image


Top
 Profile  
 
 Post subject: Re: EarthNight
PostPosted: Thu, 05-12-13, 8:23 GMT 
Offline
User avatar

Joined: Tue, 04-09-07, 21:55 GMT
Posts: 773
Location: N 42.38846 W 83.45456
not forgotten , just on the back burner
working shots before i colorize

Still in 32bit float format
In the first one you really can make out the DMZ and Seoul and the fishing fleets
Image Image Image

i am also planning on posting a 32 bit float that is histogram matched to the 2006 calibrated


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 21 posts ]  Go to page 1, 2  Next

All times are UTC


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group