3D Scan of a rabbit statue (via Entire Shape Acquisition Technique Using Multiple Projectors and Cameras with Parallel Pattern Projection)
Scanning a 3D object is nothing new. Scanning while that object is moving is a task still in its infancy. Some first steps may have been taken. Researchers from AIST (Advanced Industrial Science and Technology) in Japan, with cooperation from researchers at Kagoshima and Hiroshima Universities, have developed a method of accurately measuring 3D objects using a series of projectors and a camera.
Sure there are systems capable of doing the same thing, such as Microsoft’s Kinect and the ASUS Xtion, but they have difficulty when the object being tracked moves rapidly or changes shape. The research team’s system uses a projector (or a series of LCD video projectors) to project a specially designed ‘wavy-light pattern’ onto an object. The resulting pattern images are then captured by a Point Grey Research Flea2 camera and processed using a computer to accurately measure the object using a single image. The team was also successful using 6 high-speed cameras along 6 projectors that can measure and track an object (triangulation) moving in real time. The projectors display a set of parallel lines only, in two different colors, to be able to easily analyze the image to gain accurate measurement all the way down to 1-2mm. This makes the system capable of identifying and measuring wrinkles in clothing and even the fine details (subtle curves) of human hands.
Concept and process (via Entire Shape Acquisition Technique Using Multiple Projectors and Cameras with Parallel Pattern Projection)
The researchers state that their system/method could be used in the same fashion as the Kinect or Xtion in relation to multimedia, VR and games but has the potential of moving beyond those in scenarios involving how athletes move on a playing field that could result in reduced injuries by analyzing the player’s movements. It could also be implemented for use with stereo microscopes to measure and map microscopic images, from bacteria or viruses, to garner a better understanding of how they work or evolve. Eventually the team hopes to replace the visible pattern from the projectors with infrared versions that do not interfere with the visible appearances of objects and also allow them to capture dynamic scenes (changing or moving) and visible textures at the same time.
Diginfo report:
Cabe
