Calculating the orbit of the satellite from google Earth's images

[MrOmnos]

Hi, I was viewing Empire state building on google earth and google earth has really nice high resolution images of the building. Those pictures seems to be taken at an angle and from the certain side of the building. So,  would it be possible to predict the location in the sky where the satellite was when it took the picture from those maps images?
Just a thought.

[Someone]

The close images with that sort of detail are usually aerial photography from fixed wing aircraft, not satellite imagery.

[Ian.M]

Even if they are satellite photos, a photo of one location + am accurate 3D model of the location to match against would only give you azimuth and elevation angles of the satellite from that location.   You'd need two such photos of widely separated locations taken at the same time to determine one point in its orbit and three such points, widely separated,  to unambiguously calculate the orbit.

[Brumby]

Just how accurately did you want to calculate the orbit?

Certainly, you could get a reasonable idea of position from two images taken at the same time from the same point where the identifiable ground features are widely separated.  A third such image would improve the accuracy, but this defines only a point in space.  There would be many orbits that could pass through this one point.

You would need another set of images taken at a different time to define a second point before you could even begin to reduce the orbital possibilities.  Ideally, you would also have the time difference between the two sets so you could calculate velocity and start deriving orbital parameters.

If you did not have the time difference, you might be able to work through the orbital mechanics to derive a solution - but you would have to know how many orbits separated the sets of images.  Were they from the same orbit - or did the second set of images occur four orbits later?  The vectors are wildly different.


It's an interesting question - but I doubt Google Earth will have enough information to give you much of an answer... and all you need is one of these to really muddy the waters....

The close images with that sort of detail are usually aerial photography from fixed wing aircraft, not satellite imagery.

[SeanB]

The satellite images will invariably be a result of a satellite in a near circular polar orbit, and at such an angle that pictures of any location are always taken nearly at local noon. That way the shadows are smallest, and there is best detail. The satellite will be orbiting and scanning an area almost directly below it, any shadows will be on the pole side of the building, as it will be imaging a scan line from directly above. Errors will only occur for objects off to the side of the scanned track, which is typically a few kilometres wide, depending on the desired image resolution it can vary from only 100km wide to over 2000km with doing wide spectrum scans. The image is not a picture per se, just a series of scan lines, aligned to look like an image when properly aligned in software. No picture sensor there in most cases, you have either a scanning spinning mirror scanning the area onto a photodetector ( or a set of them) or a line scanner looking through a mask.