Pittsburgh (PA) - Scientists at Carnegie Mellon have taken the standard digital camera and expanded it into a panoramic image capture tool.  By using their assisted robotics and related software, the average digital camera can be transformed into a tool used to create breathtaking, navigable panoramas.

The technology works similarly to Google Earth.  You can move around in the image and the closer you zoom the more detail is shown.  To create the effect, Carnegie Mellon has partnered with Charmed Labs to capture a series of images by a special robotic arm which holds the camera.  Standard photographs from whatever resolution digital camera is used are then blended together through Carnegie Mellon's composite software.  The many images are merged into an interactive panorama.  From there, the compiled image data can be deployed on a web-page, on your desktop or anywhere else that digital media can be manipulated with flash.

The prototype robot connects to standard digital cameras. It then moves the camera around in a series of controlled shots, creating the high definition panoramic view. Software later combines all of the images into the Google Earth-like navigable database.

Another special camera is available which can take time lapse captures.

Carnegie Mellon has also created a website called Gigapan where the tool can be seen in action.  Each time the page is refreshed another image is shown.  By zooming in using the standard Google Earth-like controls on the left-side, more detail is available.

In the following series taken from a lake vista, I was able to zoom all the way out, then in on the water, then to a section of rocks, then down to individual rocks.  The original camera shots incorporated a high-enough detail making the full zoom-in possible, yet for the purpose of display only the visually perceptible level of detail was included at any given zoom.

Full vista, zoomed all the way out.

Closer zoom in on the water.

Closer still zoom in on the rocks.

And finally a full zoom in on individual rocks. Note that this full zoom came from the original source image. Using traditional zooming abilities on photographs would not allow such high definition.

Carnegie Mellon has also introduced several other image tools in recent months.  They appear to have a very active scientific research lab working in the area of image processing.  These new tools include a graphical search for images without appended meta tags through a type of image recognition, and a new process they developed called photo-swapping which allows a large online database to be queried for images which can be "photoswapped" into the destination image.  This technology seems to be the most desirable as there are countless organizations which routinely take group photos.  By leveraging the benefits of high-definition digital image capture abilities, more detail than ever can be maintained and navigated with a common set of tools like Google Earth.

Addendum:  For everyone posting comments that this is not a new idea I think we would all agree.  I even did this before digital cameras existed by taping picture together in my photo album.  So, it's definitely not new.  However, the idea here allows a relative novice to take their existing digital camera, mount it on the robotic device and, using the provided software, immediately create professional quality panoramic vistas.  The robotic arm carries out all of the hard work of getting everything captured just right, and the software assembles it in such a way that it's scalable.  And the display technology allows the Google Earth-like interface.  It just makes it easier, and all wrapped up in one package.

Comments (14)

Show/Hide comment form