search

cameronleger / goes16-background

goes16-background is a Python 3 script that fetches near-realtime (~15 minutes delayed) Full Disk Natural Color image of Earth as its taken by GOES-16 and sets it as your desktop background.
MIT License
34 stars 9 forks source link

Linking with himawaripy? #1

Open boramalper opened 6 years ago

boramalper commented 6 years ago

Hey @cameronleger

Great project, congrats! Would you like me to link goes16-background in himawaripy's README so that people can learn about alternatives? I'd appreciate if you do the same. =) Also I think there was a macOS specific fork but I can't remember the details right now, I'll ask the author of that fork as well to join us.

Good luck and let me know if there is anything I can possibly help with.

P.S. your README needs reformatting.

cameronleger commented 6 years ago

Hi Bora, I appreciate the outreach and any coverage you would give this project. Your Attributions was at the very bottom of the README, and I've put you and himawaripy there. I'm open to adding it to the top as well, to say something more like "looking for a western hemisphere version?"

My initial search of data lead me simply to NOAA's "GeoColor" version. I did this around noon so it was fully lit. After a few days of it I actually don't like it! The infrared bands are weird at night and I'd prefer to not see the clouds then as well. So, I scavenged some more and found a secondary source of data with "natural color" that I'll use instead; I can update the README once I do that.

That brings me to another point, though. I don't think I published it, but I did modify himawaripy to use a natural color output from CIRA so that the landmasses were green instead of red. The script was quite a departure because I was scraping a webpage instead of a JSON API, but at this point I've found a JSON API for it and it's not too dissimilar than what you're using now, if you're interested on updating himawaripy. Here's the example images.

boramalper commented 6 years ago

Your Attributions was at the very bottom of the README, and I've put you and himawaripy there.

I must have missed that! Sorry, there is no need for any further changes.

I was scraping a webpage instead of a JSON API, but at this point I've found a JSON API for it and it's not too dissimilar than what you're using now, if you're interested on updating himawaripy. Here's the example images.

The link you posted is the imagery taken by himawari-8, or am I confused? Which satellite are you talking about?

cameronleger commented 6 years ago

himawaripy pulls images from Himawari-8 for a result like this: https://himawari8-dl.nict.go.jp/himawari8/img/D531106/1d/550/2015/07/07/015000_0_0.png

Numerous alternative sources provide Himawari-8 images in the original bands as well as composite images. One of the best that I've found comes from CIRA/RAAMB at Colorado State University, where they create a "natural color" image which tries to mimic what a human would see looking from the same position. Notice the "correctly" colored landmasses: http://rammb.cira.colostate.edu/ramsdis/online/images/himawari-8/full_disk_ahi_natural_color/full_disk_ahi_natural_color_20171228070000.jpg

One downside is that the data is not always updated on time and can sometimes skip areas of time. This could leave you with an old desktop image for a few hours in the worst cases.

boramalper commented 6 years ago

One of the best that I've found comes from CIRA/RAAMB at Colorado State University, where they create a "natural color" image which tries to mimic what a human would see looking from the same position.

Hmm, I think it looks weirder to me, especially the clouds with heavy turquoise tint.

One downside is that the data is not always updated on time and can sometimes skip areas of time.

This is quite problematic I think. I mean, not that it cannot be handled, but it can be annoying to see the same image for a few hours, as you have already remarked.

What about GOES 16? How frequently is it updated and how do you fetch its images?

cameronleger commented 6 years ago

I understand your differences. There are other things I wish would be improved in the "natural color" image, specifically the nice blue fresnel-like effect that appears in the images you fetch from Himawari-8 which nicely highlight the edges of Earth.

GOES-16 provides full-disk images every 15 minutes. Originally, I was fetching an all-in-one "GeoColor" image. I created this during the daytime and it looked alright, but at night I really didn't like the imagery it provided; it faked city lights and used an infrared band to highlight oceans and clouds while landmasses were black. Now, I'm using very similar imagery to what I linked you to for Himawari-8. Here's a link to the "slider" configured to how I fetch it. The process is extremeley similar to fetching and creating the himawaripy images; notable differences include reversed x/y coordinates, different JSON structures, and different "levels."

Unfortunately, they don't yet have a nice archive page of previous imagery, so I cannot point out exactly how often it's updated or when gaps have occurred. From my anecdotal experience since the creation of this app, I've noticed a large gap once that lasted nearly a day, but otherwise it was updated frequently.