At 1 px/m², the Earth's surface is 5×10^14 pixels, so with the full constellation of 10,000 satellites and 1 fps that'd be 5×10^10 pixels/satellite/second (about a terabit/satellite/second uncompressed). If you can shoot 30 images per second, you'd need a 1.6 gigapixel camera.

It's technically possible but not feasible.

It would be quite useful at lower resolutions too, I think.

Edit: The uncompressed bitrate comment is a bit like saying 1080p network video streaming will take decades to become feasible, because the incompressed bitrate is 1.485 gigabit/s...

The obvious first bandwidth issue here is through the optical sensor, which is definitely pre-compression, no?

"It's technically possible but not feasible."

Not to mention that earth imaging satellites require approval by the government and, I believe, compliance with international regulations.

Why downvote?

It's really frustrating to see people downvote a factual statement (as opposed to opinion) without giving any sort of correction to that information. If I'm wrong - show me! I don't need this sort of crap after working all day at a job I hate where my voice is basically worthless and a legitimate debate can't take place.

They could do something like Sentinel's 10m/px, but looking at those images [1] I don't see the advantage of updating that more often than the once every few days we already have from Sentinel.

To make near-realtime interesting you need something closer to 1m/px where you can clearly make out cars. But at that point the optics and camera take more mass than a normal star link satellite. They would become earth observation satellites with an internet uplink, not the other way around.

1: https://apps.sentinel-hub.com/sentinel-playground/

Yeah, I guess it would require a breakthrough in optics/lenses to become feasible. Sentinel-2 is ~1000 kg, and each Starlink satellite is approx 250 kg.