Three-dimensional (3D) computer vision is a method of using cameras that allows computers to emulate human vision to build a 3D image. With 3D computer vision, a computer uses two cameras at once — just like a person uses two eyes — to build an image with depth. Aside from its use in creating 3D images and movies with recording devices, 3D computer vision also is used frequently with robotics, allowing robots to capture true 3D environments. One of the major problems in developing this system was ensuring that the cameras were aligned correctly, but many systems have perfected this technique. This method also makes 3D technology cheaper for the consumer market, because expensive image processors are not required to build the 3D image.
For 3D computer vision to work, the computer needs to use two different cameras the way people use two eyes. Both cameras record or capture an environment from different angles, allowing the computer to use an algorithm to blend the images and form real-life depth. Computers also are able to capture real-time 3D images, without the need for much processing between the capture and 3D building. This makes 3D computer vision useful for the gaming, movie and recording markets.
Aside from using 3D computer vision to make images and movies, this method also is often used in robotics, especially with robots made to move around and interact with an environment. By using the two cameras, the robot is able to understand the depth of an environment, making it more adept at working with other objects and overcoming physical obstacles such as gaps and bumps. Robotic movement also is smoother because of this understanding of depth.
The major problem in creating 3D computer vision was aligning the two cameras so they would work like eyes. Many of the initial systems using this technology could not get the cameras aligned, so images came out blurred or combined in incoherent ways. As of 2011, many systems have overcome this problem and some are available to consumers.
Before 3D computer vision, there were 3D image processors that could perform the same task of taking images and combining them to form depth. The major problem with this technique is that image processors are expensive, making them largely inaccessible for the consumer market. Cost is not as much of an issue for 3D computer vision, because the process of combining the images is rather simple. This allows the consumer market to enjoy 3D technology without a large price tag.
