November 7, 2012 at 21:46Kinect will be able to model objects
The team of engineers involved in the Kinect project is creating a new and spectacular feature codenamed Kinect Fusion, which came from Microsoft Research laboratories to real development, which will soon be included in the SDK and which will allow you to simulate 3D objects by simply placing the object in the field of view of the device.
In fact, this is about the following: the owner of Kinect, wanting to get a virtual copy of anything, starting from simple objects like a ball and ending with objects with a complex shape, must place this object in front of Kinect, which will scan it and place it in 3D space. At the same time, nothing will prevent you from scanning at once all the objects surrounding the console like a room; at the same time, if the Kinect sensors “make a mistake”, then the simulated object can be rotated before the console, and the model will be refined.
Technically, we are talking about collecting information from sensors and building a highly detailed three-dimensional map of an object or environment. It is also said that it will not take a lot of time to create such a map - a “one pass” will be enough even from a point from which the full perspective of the simulated object is not visible.
Kinect Fusion will be used for serious applications such as three-dimensional modeling, augmented reality and 3D printing, and, of course, realistic three-dimensional games. Moreover, if we recall the price of 3D laser scanners designed for approximately the same purposes - the creation of three-dimensional models of objects - then the price of Kinect with similar functionality will, of course, be much more profitable.
[ Source ]
In fact, this is about the following: the owner of Kinect, wanting to get a virtual copy of anything, starting from simple objects like a ball and ending with objects with a complex shape, must place this object in front of Kinect, which will scan it and place it in 3D space. At the same time, nothing will prevent you from scanning at once all the objects surrounding the console like a room; at the same time, if the Kinect sensors “make a mistake”, then the simulated object can be rotated before the console, and the model will be refined.
Technically, we are talking about collecting information from sensors and building a highly detailed three-dimensional map of an object or environment. It is also said that it will not take a lot of time to create such a map - a “one pass” will be enough even from a point from which the full perspective of the simulated object is not visible.
Kinect Fusion will be used for serious applications such as three-dimensional modeling, augmented reality and 3D printing, and, of course, realistic three-dimensional games. Moreover, if we recall the price of 3D laser scanners designed for approximately the same purposes - the creation of three-dimensional models of objects - then the price of Kinect with similar functionality will, of course, be much more profitable.
[ Source ]