https://app.skyfi.com/explore is much better than the sample preview at https://app.skyfi.com/sample-preview which is weird cause they claim that the sample preview is fully processed and better.

But it looks like explore is based on google earth. So the free preview is better than the paid thing ?

A lot of Google Maps "satellite" imagery is actually aerial imagery, not satellite. Getting that kind of detail and resolution with satellite photos is extremely hard / expensive.

If you want up to date imagery, you could certainly choose to task an airplane yourself, but that is going to cost a lot more than what SkyFi is charging for one-off satellite images.

So, no, the free preview is not better than the paid thing, and the reason they're using Google Maps is clearly to help you precisely mark the area that you want them to capture.

Did you zoom in? When I zoom in on one of the major intersections and look at the cars, it looks about right. Half-meter resolution means that each car should be several blurry pixels wide, and that's what I get.

I zoomed in, was also disappointed. Don't know the use case for these photos, but when you compare it to the quality of say a drone photo, SkyFi is nowhere near as good.

I'm sure there must be a market for these photos, but for most people I think a drone is probably better and more cost effective.

> but when you compare it to the quality of say a drone photo

Comparing a satellite/plane photo to a drone is apples to oranges.

There’s no way to scale a business in providing global drone level coverage.

Now there are services that fly planes with high res imagery that can get down to ~20cm. And even these business are super difficult to scale.

we have drone partnerships, airplane, stratospheric balloons, etc and are just STARTING with satellites. Drone imagery is better but a problem of scale but we are trying to solve that. Think of it not as photos but more so of data....we could never list the complete use cases here because there are so many

Yes, an image that's taken from three to four orders of magnitude further away is much worse than one taken from close up. Unless you spend a lot on the optics and make the satellite really big. But while the NRO is busy pointing space telescopes at earth, the private sector has so far been more interested in using these images to estimate crop yields over entire countries, or look at how full parking lots of certain companies are. For those use cases it's more useful to launch more lower-cost satellites to get more coverage more frequently, at the expense of quality.

When I zoom into my property, a Google logo is display. So I assume they are sourcing from them.

No, that's not how this works. They're not giving away imagery for every address for free. They have some samples provided, which your address probably isn't in.

They do use Google Maps' satellite view to help you mark the correct area for them to capture.

The explorer is actually Google Maps. It looks like they are using it for you to have a tool to select the area that interests you.

Maybe there should be a disclaimer.

That's about what you get with a satellite. What maps like Google contain is aerial imagery.

Here's [0] an interesting What If (XKCD) which deals with the resolution of the Hubble Space telescope if it were pointed at the earth.

[0] https://what-if.xkcd.com/32/

I'm curious for a re-write of this for JWST

It looks about right You can plainly see the 5 yd line markers on House Field (in Austin, TX) on the image. Those lines are at most 15cm wide - enough to seriously desaturate the green in any pixel that contains a line, but nowhere near enough to show as a sharp line.

The 50cm resolution is much better than what I expected (you can clearly see lines that are much less than 50 cm wide), but the 75 cm resolution is much, much worse than the 50cm one. Is it possible that some of the "50 cm" imagery is actually much better than 50 cm (which would defeat the purpose of a sample)?

I'm also not sure what's up with the difference between 50 and 75cm. But being able to see some features that are less than 50cm wide is what you should actually expect, even though it is a bit counterintuitive. Think of it like this: each pixel is the average color of a 50x50cm square. Now a square of black asphalt will have a very dark average. But square of asphalt with a 10cm wide white line running through it will be about 10/50 = 20% lighter. Easily enough to be visible. And many road markings are wider than that.