Rainbow plane warps in from gay dimension Further proof that Obama's America really is a much, much lovelier place comes in the form of another Google Maps polychromatic plane sighting, following the discovery last year of El Prez's rainbow stealth aircraft. In this case, the vehicle in question is warping in over South Carolina, having presumably travelled at light …
It's certainly not chomatic aberration, otherwise it would have been a true rainbow and not just 3 distinctive pictures. (Unless of course the sun has been secretly replaced by a set of 3 monochromatic light sources)
What I suspect is that their camera first shoots a blurry blue, then a green, then a red frame, before shooting a high resolution monochome one. Those 4 pictures combined will give you something like this.
There's one of them hovering over a hill outside - unfortunately still cloaked and staying that way. If you look at it directly, you won't see it unless you know precisely what it is. Your only hope is to catch it by surprise out of the corner of your eye. Sometimes jumping up and turning mid-air can help...
I looked up the GeoEye system. My question is why the separation when the system is designed to be able to take pictures from an airplane with turbulence. Is it the speed of the satellite? Seems odd though with the plane separating in directly down it's flight path.
http://www.geoeye.com/CorpSite/assets/docs/mjharden/MJH_Digital_Aerial_Imaging.pdf
I assume that it's because the camera is capturing the red, green and blue elements from its CCD at slightly different moments. The grounds is unaffected (probably some correction happening), but because the aircraft is moving fast relative the the ground it has moved noticeably between the blue, green and red snaps. Not sure why they'd do this (usually all three are captured simultaneously) - maybe some quirk of the specialised, high-res camera they were using?
Might also have something to do with the optics - someone mentioned chromatic aberration which sounds like a nice technical word and so may have something to do with it!
RGB is the Additive Color system being employed by the satellite's imaging hardware. It's used on your computer monitor and on projectors. 100% of those 3 colors projected by light add up to WHITE. (Subtractive Color is used in printing and painting - red, yellow & blue)
So since the satellite probably is collecting light back one-at-a-time in RGB (red, green and blue) and com-positing (combining) the image into a photo - and there looks to be a 4th image enhancement channel as well - and since the plane is traveling at a relatively fast speed - the RGB+ shots do not combine as a stationary (not moving) object would due to the fact that the camera couldn't capture the object at the same time in space.
So if you actually knew what plane that was and its actual size, you could determine from the image how fast he plane probably was going and how fast the cameras on the satellite are taking the shots from this photo.
The camera takes (collects from light coming back to it) 4 consecutive shots in Blue, Green, Red and image enhancement, and combines them. Each are shot separately, micro seconds apart. A stationary object will compose just fine, but an object moving at a higher velocity will end up looking like this jet.
Notice that the fuselage composites rather well. http://maps.google.com/maps?hl=en&tab=wl
Oh, and the Enola Gay (plane that dropped the A-bomb on Hiroshima) and this plane flying over Enola Dr. is really rather funny.
As mentioned above, it has much to do with the construction of the camera involved.
Consumer digital cameras (except for Sigma's Foveon) employ an interleaved (Bayer) color pixel matrix, which requires extra processing to create a full color image, that doesn't in fact have quite the effective resolution touted by its manufacturer.
Foveon and most high-tech/high res cameras use separate full resolution sensors for each color, so the resulting image is just a composition of the separate color images on top of each other.
The disadvantage of this process is the time delay between the capture of each color, which manifests itself as "rainbow patterns" in moving objects.
The camera involved in this image indeed has a very short delay between the color separations, so the effect doesn't show up very often, but an airplane moving at 800kph is just too big a challenge.
PS: the separate sensor idea is simplified to prove a point.
As Nathan Brown says, the camera uses multiple sensors - probably in this case three with colour filters and one without. These are obviously not in the same place and so would otherwise give slight registration errors between the images. So the camera system includes compensation, both for this and for any difference in timing (ie taking into account the movement of the aircraft carrying the camera) - it is likely that the sensors are not read at the same time, but one after the other for economy of hardware. The effect of the latter is that any moving object will appear in different places in the pictures.
But the aircraft shown is moving "fairly quickly", so it looks like the blue image was read first, then the green image, then the red one, and finally a "panchromatic" (ie sensitive across the whole spectrum). Between reading the blue and green cameras, the aircraft has moved a bit, then by the time the red is read, the aircraft has moved a bit more, and finally it's moved even more when the last image is read.
The computer running the system will adjust the images to take account of the movement of the camera, so the ground will fit together perfectly. What it can't do is correct for the fast moving object in the middle of the picture !
The same process, and same issues can occur in lower cost systems that use one sensor and multiple colour wheels - ie it applies the blue filter and takes an image, applies the green filter and takes an image, and so on. During each delay, the scenery changes, and a fast moving object will appear as shown.
Since the RGB are almost aligned along the length of the fuselage (5 deg dif?), the two craft were probably running in almost exact opposite direction to each other, multiplying the speed difference.
Also, a camera that needs to combine these images will be calibrating itself to the altitude above the surface it is imaging, so something 1-10,000 feet short of the expected focus will also screw up the timing as well.
Reminds me of trying to take color shots of nice sunsets using an Amiga, a DigiView Gold box, a consumer-grade camcorder, and an R-G-B Color splitter. Even on a relatively calm day it was, due to the capture speed, completely impossible to have all 3 passes line up properly.
Ooer, I'm dating myself a bit there . . .
The best speculative theories I can come up with are as effects of speed (sequential imaging), or height (multiple imagers).
Since cars on motorways travel at a significant portion of the plane's speed, they should be blurry, but they're not. Buildings are not a significant portion of the cruising height of planes like that, so that points to a satellite that uses many imaging systems, one for each colour. Clearly this is silly, because the satellite would have to have RGB systems metres apart at least to create this (the plane is still some way away).
The best one I've got is that, as things get closer to the camera, their speed is accentuated. If we assume that the effect is based on the satellite sequentially imaging channels, then it must compensate in software for its own path across the surface of the earth. Doing this would result in any close objects having accentuated (and inaccurately compensated) speed effects. This would also explain why the 'grey' channel causes the fourth ghost image.
Satelites and a lot of aircraft based cameras are pushbroom.
They have a 1-d line camera with a row of red then a row of green then blue pixels + often an infrared line as well.
The output from the 3 or four lines is combined (taking into account how far the cameras has moved along the track) - if something moves much faster than the carrying aircraft or satelite track in the same direction then it could move between the lines of pixels.
This post has been deleted by its author
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
