Re: New version EUMETCastView 1.1.8


Christian Peters
 

Hugo,

that's really interesting!

So no change necessary in your code-all well but I think you already
implemented some workaround...

Thank you very much.

Regards,

Christian

Am 18.12.16 um 14:44 schrieb hvanruys@... [MSG-1]:

Christian,

I discovered the cause of the artefact. It was an anomaly in the
geolocation table geo_coordinates.nc of this particular segment.


Debug Debug: "-------------> i = 1830 j = 2482 latpos1 = 46.6301
lonpos1 = 8.11723"
Debug Debug: "-------------> i = 1830 j = 2483 latpos1 = 46.6293
lonpos1 = 8.12213"
Debug Debug: "-------------> i = 1830 j = 2484 latpos1 = 46.6287
lonpos1 = 8.12662"
Debug Debug: "-------------> i = 1830 j = 2485 latpos1 = 46.628
lonpos1 = 8.13099"
Debug Debug: "-------------> i = 1831 j = 2482 latpos1 = 46.6276
lonpos1 = 8.11564"
Debug Debug: "-------------> i = 1831 j = 2483 latpos1 = 47.0164
lonpos1 = 5.30587"
Debug Debug: "-------------> i = 1831 j = 2484 latpos1 = 46.6262
lonpos1 = 8.12534"
Debug Debug: "-------------> i = 1831 j = 2485 latpos1 = 46.6255
lonpos1 = 8.12981"
Debug Debug: "-------------> i = 1832 j = 2482 latpos1 = 46.6252
lonpos1 = 8.11419"
Debug Debug: "-------------> i = 1832 j = 2483 latpos1 = 46.6244
lonpos1 = 8.11935"
Debug Debug: "-------------> i = 1832 j = 2484 latpos1 = 46.6237
lonpos1 = 8.12405"
Debug Debug: "-------------> i = 1832 j = 2485 latpos1 = 46.623
lonpos1 = 8.12858"
Debug Debug: "-------------> i = 1833 j = 2482 latpos1 = 46.6227
lonpos1 = 8.11276"
Debug Debug: "-------------> i = 1833 j = 2483 latpos1 = 46.6219
lonpos1 = 8.11789"
Debug Debug: "-------------> i = 1833 j = 2484 latpos1 = 46.6212
lonpos1 = 8.12276"
Debug Debug: "-------------> i = 1833 j = 2485 latpos1 = 46.6205
lonpos1 = 8.1273"


For linenumber i =1831 an x-position j = 2483 the geolocation latitude
an longitude are wrong. My smoothing routine did not handle this
anomaly very well. I took some precautions in the routine to avoid
that in the future.
That was an interesting problem !


Kind regards,


Hugo






---In MSG-1@..., <hctpeters@...> wrote :

Hugo,

good to know you could reproduce that artifact.

Thanks for enlighten me concerning extension take with no extension.
So default is .jpg ...that's good to know.

Have a nice weekend.

Regards,

Christian

Am 18.12.16 um 11:33 schrieb hvanruys@... mailto:hvanruys@... [MSG-1]:

Hi Christian,

I could reproduce the artefact. Very strange. It is something the
smoothing algorithm. I will look into that.
When you do not give an extension (jpg or png), the program will take
the jpg as default.


Kind regards,
Hugo


---In MSG-1@... mailto:MSG-1@...,
<hctpeters@...> wrote :

Hugo,

two little things I mentioned:

- I combined two segments EFR S3A from 17-12-2016 and did a projection.
I got an artifact inside the picture (see link)?

https://dl.dropboxusercontent.com/u/22846407/artifact_combining_two%20segments_S3A_145_17_12_2016.jpg
https://dl.dropboxusercontent.com/u/22846407/artifact_combining_two%20segments_S3A_145_17_12_2016.jpg

https://dl.dropboxusercontent.com/u/22846407/artifact_combining_two%20segments_S3A_145_17_12_2016.jpg
https://dl.dropboxusercontent.com/u/22846407/artifact_combining_two%20segments_S3A_145_17_12_2016.jpg

- If I save my picture and .png is preselected filling in a filname like
'test' ended up with a .jpg picture?
If I choose 'test.png' then the png is stored.


Regards,

Christian


Am 15.12.16 um 13:31 schrieb hvanruys@... mailto:hvanruys@... [MSG-1]:

Hello,


A new version of EUMETCastView is released. You can again just copy
the exe file, but you also have to copy the new FreeImage.dll to this
directory. In case you do not have vcomp120.dll it is also included in
the ZIP file.

The scale_factor and add_offset are now used to calculate the
radiances. Also a bug in the GVP projection for negative latitudes was
fixed.
Histogram matching is added for OLCI images.
None 95% Linear mapping of 95% of the pixels to 0 - 255. The lower
2.5% of the pixel values are set to 0 and the higher 2.5% of the pixel
values are set to 255. None 100% Linear mapping of 100% of the pixels
to 0 - 255 Equalize
Histogram equalization of the complete or projected image. The LUT is
calculated over the complete input image.
Equalize Projection (only for projections)
Histogram equalization of the projected image. The LUT is calculated
only over the pixels of the projected image.
There is also a radiobutton where you can combine the histogram
matching with normalized values. These values are calculated by
dividing the radiances by the cosine of the Sun Zenith Angle.
For OLCI images you can now save the image as a 48bit RGB PNG file.

A "Whats This ?" button was added to the Toolbar. It provides a simple
description of any widget, i.e. answering the question "What's This?".
Pushing this button and hovering over any widget in the Toolbox
changes the cursor in a question mark, and by clicking gives a short
explanation of the widget.
For the moment only a few of the new widgets are provided with help
texts.


Kind regards,


Hugo




[Non-text portions of this message have been removed]



[Non-text portions of this message have been removed]



[Non-text portions of this message have been removed]



[Non-text portions of this message have been removed]






[Non-text portions of this message have been removed]



[Non-text portions of this message have been removed]

Join MSG-1@groups.io to automatically receive all group messages.