NGC Thumbnails

[Jasem Mutlaq]

Hi John,

The RGB ones have a few nice images, but most images have a lot of
artifacts, so I think it's better to stay with your initial monochrome
version for now. Thanks for the awesome contribution John! I think it is
great as it is now.

Regards,
Jasem


On Wed, Apr 19, 2017 at 7:25 PM, John S. <myxlflik at hotmail.com> wrote:

> All,
>
>
> I tweaked the settings as much as I could and I did a run of the NGC
> objects to see how many of them came out usable.  I used data from the
> DSS1R for RED, DSS for GREEN and DSS1B for BLUE.  I used the same sizing
> parameters for the original monochrome pulls.
>
>
> I am attaching a link to the onedrive shared file for you guys to pull and
> review:
>
> https://1drv.ms/u/s!AnSnRFrNRtfXviM4XnBnN9neLucR
>
>
> You will probably notice quite a few of the images have a yellow or green
> overall appearance.  I am not 100% sure if that is because of missing data
> or the histeq method I chose this time.  Probably a bit of both.
>
> Some of the images appear to be quite usable like the attached NGC1432 and
> NGC2024.  That being said, many of the objects came out like NGC10, NGC40
> and NGC56.  There are also ones that have significant source data issues
> like NGC101.
>
>
> My question to the group is this, Do we feel there is enough good picture
> data like 1432 and 2024 to warrant the time and effort to comb through the
> 7840 objects to get the color objects out and include them in the mix.
> Once we identify the 'good' data images, I have to rename and convert them
> to PNG for inclusion.  I would also intend to split up the excel
> spreadsheet so that the ones we deemed worthy would be pulled in RGB if we
> ever needed to pull them again.  The remaining black and white images would
> be either candidates for further experimentation to get the color data
> right in them or they would be left as monochrome in the thumbnails
> collection.
>
>
> Please weigh in on this and also let me know if you would be willing to QA
> these images and give me a listing of the good ones you find.  This is a
> fairly time consuming effort and I would like some help as well as guidance
> on what others feel are decent images or not.
>
>
> John
>
>
>
> ------------------------------
> *From:* John S. <myxlflik at hotmail.com>
> *Sent:* Tuesday, April 18, 2017 3:01 PM
>
> *To:* Jasem Mutlaq; KStars Development Mailing List
> *Subject:* Re: NGC Thumbnails
>
>
> Jasem,
>
>
> You can use my last name (Sadler) for the credit on the packages.  I
> played around with the data some more and have gotten a few images to come
> out acceptably, but most of completely out of balance when you convert them
> to RGB.  I think I found out why.  One of my underlying assumptions about
> the DSS data appears to be incorrect.  Although the original DSS survey is
> monochrome, it does not appear to be filtered.  It appears to be visible
> light.  Obviously when you assign full spectrum to one channel it's going
> to play all kinds of havok with the balance of the channels.  I am not
> certain if there is a practical way to deal with the data, but perhaps a
> 'green' channel can be extracted by subtracting the red and blue channels
> from the DSS green channel.  This all would require significant effort but
> may yield results and be worthwhile if we can find a way to do it in a
> batch manner.  I can certainly pull all three channels separately and start
> playing to see if I get any kind of real results.  I will let you know what
> I find out.
>
>
> John
>
>
> ------------------------------
> *From:* John S. <myxlflik at hotmail.com>
> *Sent:* Tuesday, April 18, 2017 9:27 AM
> *To:* Jasem Mutlaq; KStars Development Mailing List
> *Subject:* Re: NGC Thumbnails
>
>
> Great.  I am continuing to tweak and play with settings on the object
> pulls to see if I can get better data.  The problem I think is that they
> don't have a setting to adjust the levels and it seems like the blue
> channel is too weak overall to make for a proper color balance.  The last
> thing I want to do here is manually adjust 13000 images.  Even with a batch
> process it feels like a lot of effort.  I will keep you posted as to what I
> find.  I will start 'versioning' the tarball files as I change them so that
> you will know if it's newer than what you have now.
>
>
> ------------------------------
> *From:* Kstars-devel <kstars-devel-bounces at kde.org> on behalf of Jasem
> Mutlaq <mutlaqja at ikarustech.com>
> *Sent:* Tuesday, April 18, 2017 8:13 AM
> *To:* KStars Development Mailing List
> *Subject:* Re: NGC Thumbnails
>
> Yup we can publish an updated version if we get better data
>
> On Tue, Apr 18, 2017 at 2:59 PM, John S. <myxlflik at hotmail.com> wrote:
>
>> Oh wow.  I thought you wanted to try to work on them some more.  Assuming
>> we can get better images for  this, what will the process be to update this
>> now that it is incorporated?  Just publish a new version and have them
>> download it via that data downloads in Kstars?
>>
>>
>> ------------------------------
>> *From:* Jasem Mutlaq <mutlaqja at ikarustech.com>
>> *Sent:* Tuesday, April 18, 2017 4:38 AM
>> *To:* John S.
>> *Cc:* Hans; KStars Development Mailing List
>> *Subject:* Re: NGC Thumbnails
>>
>> I just added them to KStars. I used your name "John S." and email
>> address. Please let me know if you want to change any of those as they
>> appear in the download dialog. I'll try to add previews as well since we do
>> not have previews for _any_ download before.
>>
>> On Mon, Apr 17, 2017 at 7:44 PM, John S. <myxlflik at hotmail.com> wrote:
>>
>>> Did anyone have any further input on this?  Is it worth pursuing and
>>> adding to Kstars?
>>>
>>>
>>>
>>> ------------------------------
>>> *From:* John S. <myxlflik at hotmail.com>
>>> *Sent:* Saturday, April 15, 2017 11:00 AM
>>>
>>> *To:* Hans
>>> *Cc:* KStars Development Mailing List
>>> *Subject:* Re: NGC Thumbnails
>>>
>>>
>>> Here is a link to the current NGC file tarball and the rudimentary
>>> install and removal scripts.  This has only been tested on Linux Mint, but
>>> suspect it will be fine on all versions:
>>>
>>> https://1drv.ms/u/s!AnSnRFrNRtfXviEYxBIM2pNlmuc6
>>>
>>>
>>> Here is a link to the current IC file tarball.  Same caveats as the NGC
>>> file:
>>>
>>> https://1drv.ms/u/s!AnSnRFrNRtfXviLmrzlp6Ru6N7KZ
>>>
>>> The images are all of the DSS2R (Red Plate) survey only.  There is an
>>> ability to pull data into RGB using the DSS2R, DSS and DSS2B surveys (or
>>> really any other combination for that matter).  I have found that in nearly
>>> all cases, the colors needed subsequent adjustment in order to make the
>>> images look palatable.  I honestly felt the red channel ones and gray scale
>>> worked out just find overall.  I know there are some outright bad images in
>>> there and I haven't had the time or inclination to weed through the 13,000
>>> objects I pulled to clean it up as of yet.  I can certainly do it, but it
>>> would largely depend on how strongly this group felt about it's inclusion.
>>> I agree it should be an add-on package and not part of the base install.
>>> We will need to make sure to credit the sources of the images as well.
>>>
>>>
>>> One thing we may need to consider is directory management.  Downloading
>>> all of these images puts 13,000+ files in the $HOME/.local/share/kstars
>>> directory.  That's a pretty good chunk of files for a single directory.  I
>>> may make sense to compile these into some sort of database file or put them
>>> in a directory structure.  I am not a developer though and have NO skill
>>> when it comes to that aspect of this project.
>>>
>>>
>>> John
>>> ------------------------------
>>> *From:* John S. <myxlflik at hotmail.com>
>>> *Sent:* Saturday, April 15, 2017 10:15 AM
>>> *To:* Hans
>>> *Cc:* KStars Development Mailing List
>>> *Subject:* Re: NGC Thumbnails
>>>
>>> I started this project for my own personal gratification. I personally
>>> don't mind that some of the objects are subpar quality. Most of those
>>> objects are very interesting to me anyway. However, I would like to make
>>> sure that it meets certain quality levels. I am open to input for how to
>>> rectify the situation. I believe the limits of the survey itself will limit
>>> the outlet quality though. If you want I'll post the links to the two
>>> tarballs for the NGC and IC objects and you can download them a try for
>>> yourself.
>>>
>>> John
>>>
>>> > On Apr 15, 2017, at 9:02 AM, Hans <hans at lambermont.dyndns.org> wrote:
>>> >
>>> > Agreed, this is something very nice to have. And it is tedious work
>>> too.  For
>>> > images that don't look too good I expect we'll even get negative
>>> remarks, but
>>> > that should not inhibit this project. To me having a skimpy image is
>>> far better
>>> > than no image at all.
>>> >
>>> > Is the idea to have an image library or to dynamically download them
>>> on demand ?
>>> >
>>> > -- Hans
>>> >
>>> > Jasem Mutlaq wrote on 20170415:
>>> >
>>> >> I'm attaching a few examples so you can see why we can't use them as
>>> is in
>>> >> KStars.
>>> >>
>>> >>> On Sat, Apr 15, 2017 at 9:43 AM, Akarsh Simha <akarshsimha at gmail.com>
>>> wrote:
>>> >>>
>>> >>> Hi
>>> >>>
>>> >>> This is something very nice to have!
>>> >>>
>>> >>> A caveat to consider is the licensing for DSS imagery. They are not
>>> >>> available for commercial use, so we must ship this as an add-on
>>> through
>>> >>> KNewStuff ("Get New Data")
>>> >>>
>>> >>> Other sources of images are SDSS, the Hubble Legacy Archive (
>>> hla.stsci.edu)
>>> >>> and the recently released PanStarrs 1 survey. I think the PanStarrs 1
>>> >>> survey has a large sky coverage (possibly even all sky -- I haven't
>>> >>> checked). SDSS has a somewhat small footprint, and the Hubble only
>>> has
>>> >>> objects of interest.
>>> >>>
>>> >>> I didn't quite get what the problem involved is. The DSS is made from
>>> >>> digitizing scanned plates mostly from the Palomar Observatory and
>>> UKSTU
>>> >>> telescope. The result is that many plates have artefacts, and they
>>> are
>>> >>> completely uncalibrated. Planetary nebulae and bright emission
>>> nebulae tend
>>> >>> to be overexposed. There is no way around these problems other than
>>> going
>>> >>> to other surveys.
>>> >>>
>>> >>> Regards
>>> >>> Akarsh
>>> >>>
>>> >>>
>>> >>
>>> >>
>>> >> --
>>> >> Best Regards,
>>> >> Jasem Mutlaq
>>> >
>>> >
>>> >
>>> >
>>> >
>>>
>>
>>
>>
>> --
>> Best Regards,
>> Jasem Mutlaq
>>
>>
>
>
> --
> Best Regards,
> Jasem Mutlaq
>
>


-- 
Best Regards,
Jasem Mutlaq
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kstars-devel/attachments/20170420/0c69447b/attachment-0001.html>