undefined | Better HN

0 pointsdoersino5y ago0 comments

I've recently written a Python app that selects a random location in an area defined by a user-supplied shapefile [1], grabs corresponding aerial imagery from Google Maps, and posts it as a geotagged tweet:

https://github.com/doersino/aerialbot

I've built this tool because satellite imagery can be extremely beautiful [2], and I was looking for a way of regularly receiving high-resolution satellite views of arbitrary locations such as the center pivot irrigation farms of the American heartland [3] in my timeline. Plus, for obvious reasons, it's nice to see the world without actually having to go outside right now.

Currently, I'm running two Twitter bots based on ærialbot:

* @americasquared, which posts one randomly selected square mile of the United States every 4 hours: https://twitter.com/americasquared

* @placesfromorbit, which analogously posts a 5×5 km square anywhere in the world every 6 hours: https://twitter.com/placesfromorbit

---

[1]: https://en.wikipedia.org/wiki/Shapefile

[2]: https://earthview.withgoogle.com

[3]: http://www.thegreatamericangrid.com/archives/1441

0 comments

Wookai5y ago

Awesome project, I love it! If haven't seen it yet, Google has some hand-picked highlights of places with stunning aerial pictures: https://earthview.withgoogle.com/. Some high-level info here: https://www.blog.google/products/earth/most-stunning-images-...

EDIT: I read the readme and of course you mentioned Earth View :). Leaving the links for other people who might be interested.

m3at5y ago

Earth View indeed has some stunning aerial pictures! Thanks for leaving the link, I would have missed it otherwise :)

safog5y ago

I think you can make the images a lot more appealing by adding some automated post-processing to them.

I don't know a lot about image processing algorithms but clicking "Auto" on google photos tweaks basic stuff like exposure, contrast, highlights, shadows, vibrance etc. so the image has a lot more "punch".

doersinoOP5y ago

Good comment – ærialbot actually does increase contrast and saturation a tiny bit, but I've kept this very conservative for a couple of reasons:

* Some areas of the world are just naturally fairly flat and monochromatic, so a dynamic contrast/saturation/brightness adjustment (e.g. one that would turn the darkest pixel black, the brightest pixel white and linearly map the rest between these extremes) would not work for these areas.

* The available satellite imagery has been captured and processed in a variety of ways depending on the region, so a constant contrast/saturation/brightness adjustment might work well in some places, but overcorrect things in other places (especially urban areas in the US and Europe tend to already be fairly saturated and contrasty).

Basically, doing this well would involve a whole bunch of testing and fine-tuning. And since not even Google (the source of the imagery) seems to do this, I decided not to bother: Keeping the data basically the way I receive it is easy and "honest".

j / k navigate · click thread line to collapse