Topics

EUMETCastView v1.3.9

Hugo
 

Hi,
In this new release I included the standard EUMETSAT RGB recipes from METEOSAT/SEVIRI data as described by this document : http://www.eumetrain.org/RGBguide/recipes/recipes.pdf .
The images produced take a bit longer to calculate than normal. Some parts of the code ( calculations of brightness temperature, radiance and reflectance ) are taken from https://github.com/gmcgarragh/seviri_util made by Simon Proud.
You can download a Windows version from https://github.com/hvanruys/EUMETCastView/releases . Hugo

Douglas Deans
 

On 28/06/2018 09:28, Hugo wrote:
Hi,
In this new release I included the standard EUMETSAT RGB recipes from METEOSAT/SEVIRI data as described by this document : http://www.eumetrain.org/RGBguide/recipes/recipes.pdf .
The images produced take a bit longer to calculate than normal. Some parts of the code ( calculations of brightness temperature, radiance and reflectance ) are taken from https://github.com/gmcgarragh/seviri_util made by Simon Proud.
You can download a Windows version from https://github.com/hvanruys/EUMETCastView/releases .
Hugo

===========================================================

Many thanks Hugo.
Granddaughter here today so will give it a try tonight.

Regards,
Douglas.

Christian Peters
 

Hugo,

very nice, I tested on windows first, looks good.
Is it possible to add No.10 (HRV clouds) of the http://www.eumetrain.org/RGBguide/recipes/recipes.pdf  pdf document too?
Would be very nice.

I was wodnering about the color of the Severe Storm RGB of EMCV. It's very reddish if I compare it to the product of Eumetcat, xrit2pic or made in Pytroll!?

See pics appended.

Regards,

Christian



Am 28.06.18 um 10:28 schrieb Hugo:

Hi,
In this new release I included the standard EUMETSAT RGB recipes from METEOSAT/SEVIRI data as described by this document : http://www.eumetrain.org/RGBguide/recipes/recipes.pdf .
The images produced take a bit longer to calculate than normal. Some parts of the code ( calculations of brightness temperature, radiance and reflectance ) are taken from https://github.com/gmcgarragh/seviri_util made by Simon Proud.
You can download a Windows version from https://github.com/hvanruys/EUMETCastView/releases . Hugo

Christian Peters
 

Hugo,

even with Day Microphysics seems something wrong, see attached examples??

Regards,

Christian




Am 29.06.18 um 09:10 schrieb Christian Peters via Groups.Io:

Hugo,

very nice, I tested on windows first, looks good.
Is it possible to add No.10 (HRV clouds) of the http://www.eumetrain.org/RGBguide/recipes/recipes.pdf  pdf document too?
Would be very nice.

I was wodnering about the color of the Severe Storm RGB of EMCV. It's very reddish if I compare it to the product of Eumetcat, xrit2pic or made in Pytroll!?

See pics appended.

Regards,

Christian



Am 28.06.18 um 10:28 schrieb Hugo:
Hi,
In this new release I included the standard EUMETSAT RGB recipes from METEOSAT/SEVIRI data as described by this document : http://www.eumetrain.org/RGBguide/recipes/recipes.pdf .
The images produced take a bit longer to calculate than normal. Some parts of the code ( calculations of brightness temperature, radiance and reflectance ) are taken from https://github.com/gmcgarragh/seviri_util made by Simon Proud.
You can download a Windows version from https://github.com/hvanruys/EUMETCastView/releases . Hugo


Hugo
 

Christian, Graham,

In the 'Severe Storms' RGB recipe parameters was a small bug. This is  fixed ,and the 'red' globe is now 'purple'.
However for the Snow and Day Microphysics ( which uses the IR_039 sun reflection ) I am still searching for the reason.
There are many ways to calculate the IR_039 sun reflection. My calculations include the CO2 correction from the IR_139 band, maybe this is the reason.
I will experiment a bit more later today. 
grts,

Hugo

Ernst Lobsiger
 

Hugo,

If the Day Microphysics image from EMCV is the version at left posted by Christian it seems far better. No need to deteriorate it :-) !

Cheers,
Ernst

Christian Peters
 

Ernst,

the version with more orange-blue is the version provided by Eumetsat (even similar colors with xrit2pic and Pytroll). The version with red and cyan is the EMCV version.

(see filenames)

Regards,

Christian

Am 30.06.18 um 10:21 schrieb Ernst Lobsiger:

Hugo,

If the Day Microphysics image from EMCV is the version at left posted by Christian it seems far better. No need to deteriorate it :-) !

Cheers,
Ernst

Christian Peters
 

Hugo,

thanks for fixing it. Yes...I think there are several ways to calculate the RGBs.

Regards,

Christian

Am 30.06.18 um 09:27 schrieb Hugo:

Christian, Graham,

In the 'Severe Storms' RGB recipe parameters was a small bug. This is  fixed ,and the 'red' globe is now 'purple'.
However for the Snow and Day Microphysics ( which uses the IR_039 sun reflection ) I am still searching for the reason.
There are many ways to calculate the IR_039 sun reflection. My calculations include the CO2 correction from the IR_139 band, maybe this is the reason.
I will experiment a bit more later today. 
grts,

Hugo

Ernst Lobsiger
 

Christian and Hugo,

O.K. the the EUMETSAT version is superior then. I just looked at the thumbnails on the MSG1-list: Earlier the (reddish) EMCV Image was at left and now the (rather pale) EMCV Image is at right. Unless I open the images in IE I get just names 0,png and 1.png. So I was kind of confused. Hugo will certainly fix it :-).

Cheers,
Ernst

Simon Proud
 

Hi Hugo,
I'm glad that you're using our seviri_util code, which was also made by Greg McGarragh, not just me :-)
But can I ask that you include the copyright notice and a link to the github repo where you use our code? I see a few files in EUMETCastview that have been taken directly from seviri_util but do not respect the GNU GPL licence. To my (limited) knowledge the files from seviri_util must remain under GPL. Sorry to be a pain in the ass, but Greg and I were employed to develop that code and as it's good to ensure it's distributed in accordance with what our employer asked for. Thanks!

Regarding the RGB colour issues, I had a look at your code but there's a lot of files and it's not immediately clear where the brightness temperature and reflectance calculations are done. I see that you're using a similar method for the 3.9 channel as PyTroll does, but there are some differences. I'd suggest using exactly the same formulae as PyTroll and seeing if that helps. Their code starts is near line 80 of this: https://github.com/pytroll/mpop/blob/master/mpop/instruments/seviri.py

You could also try the method we use in our retrieval software (ORAC), which explicitly calculates the solar term. You can see it in this file, starting on line 337: https://github.com/ORAC-CC/orac/blob/master/src/FM.F90
This might be overkill for your purposes but it works nicely for us! The F0 value for SEVIRI PFM is 4.6265.

Is it possible to save images showing the raw brightness temp and reflectance values calculated by EUMETCastView? If so, I'll compile the code and compare a couple of outputs to ORAC to see if that helps narrow down the differences.

Cheers,
Simon

Hugo
 

Hi Simon,

I will include a copyright notice and a link to the github repository. The code is of course a bit changed to the Qt style, but I will include this notice in the cpp files where the radiance and reflectance are calculated.
I should also include the '
GNU GENERAL PUBLIC LICENSE version 3' in the github repository . This could be necessary now that Github is taken over by Microsoft ...
For the calculation of the IR_039 reflection I used a document from Eumetrain. (eumetrain.org/IntGuide/PowerPoints/Channels/conversion.ppt). A comparison between the Pytroll code and my C++ code should reveal some differences.
Thanks for the tip. 
If I do not find the reason for the difference  I will include a menu entry that can create a file with, for example100 x 100 brightness temps or reflectances. I will keep you informed.
Cheers,
Hugo

Simon Proud
 

Hi Hugo,
Thanks for that, much appreciated.
I have a lot of difficulty reading Danny's presentations, so wouldn't surprise me if there's an error in there somewhere! The Pytroll code looks correct to me so that's probably the best example to follow.
If it's a lot of work to update the menu then I wouldn't worry about that, if you can do a debug hack to save BTs + Refl for one recent timeslot then just send those to me and I'll have a look, probably simpler than altering the whole code.

Cheers,
Simon

On 01/07/18 12:07, Hugo wrote:
Hi Simon,

I will include a copyright notice and a link to the github repository. The code is of course a bit changed to the Qt style, but I will include this notice in the cpp files where the radiance and reflectance are calculated.
I should also include the '
GNU GENERAL PUBLIC LICENSE version 3' in the github repository . This could be necessary now that Github is taken over by Microsoft ...
For the calculation of the IR_039 reflection I used a document from Eumetrain. (eumetrain.org/IntGuide/PowerPoints/Channels/conversion.ppt). A comparison between the Pytroll code and my C++ code should reveal some differences.
Thanks for the tip. 
If I do not find the reason for the difference  I will include a menu entry that can create a file with, for example100 x 100 brightness temps or reflectances. I will keep you informed.
Cheers,
Hugo

-- 
Dr Simon R Proud
NERC Post-doctoral fellow in Aviation Meteorology,
Department of Atmospheric, Oceanic and Planetary Physics,
University of Oxford,
Clarendon Lab,
Parks Road,
OX1 3PU,
Oxford,
UK

Tel (office): +44 (0) 1865 282431
Tel (mobile): +44 (0) 7563 639685
Email: simon.proud@...
Twitter: @simon_sat

Graham Woolf
 

Hi Simon

I have been running your Sentinel script to process sentinel3 images but have run into a problem

Its these lines

    r=scipy.misc.bytescale(outimg[:,:,0],cmin=rmin,cmax=rmax)
    g=scipy.misc.bytescale(outimg[:,:,1],cmin=gmin,cmax=gmax)
    b=scipy.misc.bytescale(outimg[:,:,2],cmin=bmin,cmax=bmax)

This function is missing from the latest scipy package and I wondered if you know what to replace it with to get the script to work

I have attached the script I have been using

I would love to get this working or if you have a more up to date script would you mind sharing it

Many Thanks

Graham Wool