Bowtie deletion and RGB formula for Image bands (Suomi-NPP)

New McIDAS-V features under development. Feedback is appreciated.
Post Reply
User avatar
michaelav
Posts: 5
Joined: Fri Aug 22, 2014 2:01 pm

Bowtie deletion and RGB formula for Image bands (Suomi-NPP)

Post by michaelav »

Dear all,

I would like to display Image bands from VIIRS in McIDAS-V 1.4, unfortunately I fight black stripes. A RGB formula for VIIRS data VIIRS_M_RGB(M5,M4,M3) deals with bowtie deletions very well, however, it creates an image in worse resolution than my data are, even if I subset the required region in all selected bands (the green rectangle with shift+click). I lose that high resolution when I use Formulas -> swathToGrid as well.

I use I-band data (375m, file GIMGO-SVI01-SVI02-SVI05_npp_d20140721_t1225328_e1231132_b14147_c20140722120529348931_noaa_ops.h5 as an example) and this formula is for M-bands (750m), please, look at images. If I use Formulas –> Imagery –> Three Color (RGB) Image (Auto-scale) formula, the resolution is OK, but bowtie deletions appears. The combination of these two layers is the best, as you can see at the third image, but... I´m afraid I need some VIIRS_I_RGB(I1,I2,I5) formula instead of the VIIRS_M_RGB(M5,M4,M3)...?

Does it happen because formulas are recommended for M bands with worse resolution? How to deal with bowtie deletions in the high resolution, please?

Besides, if I understood correctly, this formula (VIIRS_M_RGB) works only with multi-banded granules, otherwise an error report shows up. This little complicates one´s work when you would like to create different RGB composites with same bands - you have to download multi-banded granules. Maybe there is another, simpler solution?

Thank you for your responses.
Attachments
overlap of these two layers
overlap of these two layers
Three Color (RGB) Image (Auto-scale) formula
Three Color (RGB) Image (Auto-scale) formula
VIIRS_M_RGB(M5,M4,M3) formula
VIIRS_M_RGB(M5,M4,M3) formula
User avatar
hproe
Posts: 504
Joined: Sat Nov 27, 2010 3:46 pm

Re: Bowtie deletion and RGB formula for Image bands (Suomi-NPP)

Post by hproe »

Hi Michaela -

I have done many RGB scenes with VIIRS' I-channels. It is realtively straightforward. It needs some clicking, though. After having plugged in the 'Miscellaneous > VIIRS Formula' and loaded the GIMGO files containing the channels used in the RGB, it goes like this in the Data Explorer window:

1. Display > Imagery > RGB Composite
2. Data Sources > Formulas > swathToGrid > Create Display
3. In Select Input widget: set res to 375 (I use 300 for convenience) and leave mode at 1.0 > OK
4. In Field Selector widget: Formulas > Three Color (RGB) Image (Auto-scale) > OK
5. In Field Selector widget: attribute appropriate channels to the colour beams (use temperature rsp reflectivity units)
5.1 do this slowly waiting for the thumbnail of each channel to appear in the Region tab
5.2 in one of the thumbnails select the region of interest by sub-tending the green box
5.3 click OK

Of course, this could be put in a script. Just have to find a volunteer.

You can also use my SEVIRI RGB formulas (Miscellaneous > RGB Eumetsat) and select appropriate VIIRS channels (mainly for M-channels, of course). The attached image uses the NCOG formula and channels SVI01/02/03.

cheers, HP
Attachments
201405011545_snpp-INCOL_eruptionMichaelSaundersIsland-SouthSandwichIslands.png
User avatar
michaelav
Posts: 5
Joined: Fri Aug 22, 2014 2:01 pm

Re: Bowtie deletion and RGB formula for Image bands (Suomi-NPP)

Post by michaelav »

Hi HP,

oooh I see, the value of "res" in Select Input widget is the thing I missed when using swathToGrid formula. Thank you for your answer!

Unfortunately, I still do a mistake somewhere else when creating a RGB image. If I follow your step 1. (choose data source and then in Field Selector Fields > IMAGE > VIIRS-I1-SDR_All/Reflectance and in Displays > Imagery > RGB Composite and Create Display), there is an error “Creating display: RGB Composite TupleType: component index out of range”. So, I go only through your steps 2. – 5.3 to create a RGB image. As a data source there is a file GIMGO-SVI01-SVI02-SVI05_npp_... and Reflectance of I01, I02 and Brightness Temperature of I05 is selected in Red, Green and Blue. No error warning appears, but image is not created and “image values must be mapped to Red, Green or Blue only” is written at the bottom of the Main Display window, even when I do your step 5. Maybe it has something common with inquiry #946? Or am I wrong somewhere else?

Thanks,
Michaela
User avatar
Rick
Posts: 404
Joined: Fri Jan 16, 2009 8:20 pm

Re: Bowtie deletion and RGB formula for Image bands (Suomi-NPP)

Post by Rick »

Hi Michaelav,

You are correct, the errors you are seeing have been written up under Inquiry 946
User avatar
hproe
Posts: 504
Joined: Sat Nov 27, 2010 3:46 pm

Re: Bowtie deletion and RGB formula for Image bands (Suomi-NPP)

Post by hproe »

Michaela and Bob -

I now remember this inquiry, instigated by me in late 2010. This problem has been cured quite some time ago, so I forgot about it. Also, I have overlooked that Michaela is running version 1.4 - not sure whether the cure is already in there. In any case, the recipy works well in recent 1.5beta2 versions. And be sure to have selected 'RGB Composite' under the Displays tab in the Data Explorer! Attached a sample where I tried to simulate Michaela's RGB with a granule at hand.

HP
Attachments
sample.png
Post Reply