India Launches World's First Stereo Imaging Satellite

sgups writes "India will tomorrow inaugurate a new launch pad at its Satish Dhawan space port near Chennai, on the south-east coast, by putting the world's first stereographic mapping satellite into orbit. The most innovative feature of the 1.6-tonne Cartosat-1 is its pair of cameras, which will give stereo images of the earth's surface that can distinguish features down to 2.5 metres across. They will directly generate three-dimensional maps that have until now been achievable only indirectly, by combining data from a large number of satellite passes over the same place. "Such a stereographic imaging system does not exist in the civil sector anywhere else," says Mr Nair, chairman of the Bangalore-based Indian Space Research Organisation (Isro). "It will give information about heights that will be very useful in applications such as planning power lines." Cartosat-1 will join what is already the world's largest cluster of non-military remote sensing satellites. Six Indian spacecraft are already observing the earth with a wide range of instruments."

Comments

What? How far apart...

[robslimo]

do the cameras have to be to get a proper parallax?

Heights?

[Neil Blender]

Uh, can't they already determine heights to high degree of accuracy with GPS or other radio wave methods? How would a picture be more accurate?

I don't get

[JohnFluxx]

I don't get. If it's not a geosync, then it's going to be moving, so they could just use 2 images from a few seconds apart to get the required images. No?

Why two cameras needed?

[Jerf]

I'm going to assume the satellite designers knew what they were doing and there is some good reason for this.

That said, given the resolution with which we know the position of a given satellite, and the low resolution of the source image in this case, what advantage does using two cameras give you, vs. taking one camera and snapping two pictures in quick succession?

Maybe they can't be snapped quickly enough? But then, you'd think the larger parallax would be helpful, not harmful.) I know consumer cameras have the basic tech now to take a snapshot of the CCD state and process it later, that tech ought to scale right with the CCD resolution, whatever it is.

Maybe this is so you can choose the parallax direction, instead of the orbit forcing your choice? Does the image processing need the parallax to show up in some particular direction relative to the light source to work?

Honest questions; knowledgeable answers appreciated. (As you can see, I can talk out of my ass too :-), I'm looking for something a little more informed.)

Re:Why two cameras needed?

[UnknowingFool]

what advantage does using two cameras give you, vs. taking one camera and snapping two pictures in quick succession?

For a parallel example, try to take a picture of the ground from a moving vehicle at 10mph. To get a decently clear and detailed picture, your film speed would have to be high. Now try to get a stereographic image of an ant hill from overhead while moving from a vehicle at 10mph. With one camera you'd have to take fast pictures and move the camera angle without motion blur. As an alternative you could take one picture, change the angle, and pass over the ant hill a second time.

Applying those techniques to satellite imagery doesn't work well. The satellite can't rotate fast enough considering how fast it is passing over a target area. Using 2 passes does work but that unfortunately expends fuel to change the position of the satellite every time. So the lifetime of the satellite is sharply reduced unless it is serviced in space. Rarely are satellites ever serviced. Those that are serviced (Hubble, ISS, etc) have to be extremely important.

Not the distance between the cameras

[Langolier]

Though it does not say this in the article, it is not the distance between the two cameras on the satellite that produces 3D imagery. The cameras will be pointed in slightly different directions, so that the image taken by one camera at time t will be paired with the image taken by the second camera at time t+x, where the satellite has traveled probably tens of miles in time x. The second camera is pointed slightly "backwards", so that it takes pictures of the same area that the first camera was shooting x seconds ago.

This is just supposition, based on the fact that two cameras on a satellite would not be far enough apart to generate parallax.

Three Corner Sat

[eggbert.net]

Over the summer and last semester I worked in a nano-satellite lab at ASU. The most recent satellite of ours that was launched was Three Corner Sat and one of its primary mission objective was sterio imaging.

http://threecornersat.jpl.nasa.gov/
http://nasa.asu.edu/
https://spacegrant.colorado.edu/tiki-index.php?pag e=3CS

Unfortunately, the two of our satellites that got launched were released at 50,000 km instead of 100,000 km so they burnt up before they could power up.

http://www.spacetoday.net/Summary/2737

Re:power lines? Riiiight.

[Lord Omlette]

Like decades if not centuries of maps can't help there.

There's a really good book called The Great Hedge of India that touches on this sort of thing. Basically, in India, everything moves around so much that maps are worthless after a couple of years.

I'm not saying you're wrong about Pakistan though, just that you're wrong about archived maps of India...