Infoseite // New 3D camera technology in development

Newsmeldung von slashCAM:


New 3D camera technology in the development of heidi - 26 Mar 2008 13:29:00
Above all, of course, for use in robotics at Stanford is a "multi-aperture camera" developed with the identification of precise spatial information should be possible. However, the procedure may be synonymous for normal photography interesting. For the camera, the pixels of an imaging sensor has been significantly shrunk to 0.7 micron, and grouped. Before each group at 256 pixels is a small lens placed - a prototype with 3 megapixel would be at 12,616 lenses. On this lens pitch is through a larger lens, the recorded picture taken, so that each pixel from at least four, slightly shifted perspectives, as appropriate (for a detailed description or see link below bei heise.
Nun lassen sich räumliche Informationen, sprich Abstände zwischen Bildinhalten perfekt auslesen and beispielsweise für 3D-Modelle verwenden; das Picture selbst wird weiterhin in 2D vorliegen. Doch nicht nur das: synonymous zur Errechnung einer "künstlichen" Tiefenunschärfe sollen diese Daten verwendet werden können (das ist synonymous der Grund weshalb wir das Project hier vorstellen). Außerdem, so die Forscher, ließen sich solche Kameras möglicherweise günstiger herstellen als herkömmliche Digitalkameras, da die Qualität des Objektives nicht mehr ausschlaggebend sei -- wohl weil die Fehler nachträglich herausgerechnet werden können.
Wann with einem funktionierendem Prototyp gerechnet werden kann, ist leider nicht bekannt, bislang wurde lediglich der Sensor gebaut (im Picture mittig unten).

This is an auto-generated entry





Antwort von Axel:

Already years ago and linked days:
http://graphics.stanford.edu/papers/lfcamera/
And it is true that with the Z-information especially the depth of field can move. A basis for generating a 3D video can still be. When Nachbarbeitung the Star Wars films for the empty canvas was the development of software needed, the Z values from the actual recordings obtained, and allowed to determine this.
You would probably first try the consumer area, 3D marketing, before you part with blurred images wörbe. Unless the installation in the Movies (synonymous will The Hobbit plastic) is already a flop.



Antwort von WoWu:

Adobe has such a model already presented, which is based on the lens before the lens combination to be and so the Z information to win.
Also, since some time the principle, already from the depth of information to win in Game programming used. (Depth-Dependent Filtering)
There is really nothing new, synonymous when a current date to prints.
It is also possible, each pixel of an CMOS synonymous future assign such information and with a filter value to be demonstrated. Depth of field is so very soon no longer be any adapters or lenses, but from the calculator. The LUT can Lenses different (and lenses) are retrieved and reproduced.
For 3-D is probably still quite a while the problem, the classical Kantenunschärfen that emerged from the angle of refraction and shows followed by realistic.



Antwort von Axel:

"WoWu" wrote: It is also possible, each pixel of an CMOS synonymous future assign such information and with a filter value to be demonstrated. Depth of field is so very soon no longer be any adapters or lenses, but from the calculator.
Very interesting. Would not it be> Off Topic <synonymous possible with individually controllable pixels, a larger dynamic range to achieve? Through a kind of combination of exposure bracketing and Pixel Shift? The DOF SonyHD HDR-1 to start the camera I then save s.zu times ;-)))



Antwort von Flo_vdH:

Well, something similar is already done by the selective gamma or different aperture values for individual pixels (clusters).
In a few years we will be surprised to see what we (then) all the hairs have gerauft. That's why I use the demonization CMOS synonymous not just because it gets a few odd effects shows. Conversely, the company synonymous still two more transistors per pic. on the sensor design and thus give him another shutter to spendier.
The current disaster, I see in the miniaturization (for mobile). Meanwhile, more photos made with mobile phones than with cameras. ... So how is this for the cameras are? Then that will be good again expensive cameras. .... But well synonymous.



Antwort von WoWu:

And I was logged in again not synonymous ... maybe a word to the dynamic range:
This is actually already on the way, derived from the photograph. Per pixel lying together a different aperture value and the sum you have a dynamic range, of which one can only dream of, where each pixel is durchgezeichnet.



Antwort von B.DeKid:

@ Wolfgang
Klär me look at what is "LUT" thank you.



Antwort von WoWu:

Look-up table: a "directory" from the values of certain settings or requirements statically or dynamically fetch.