It wasn’t that long ago that constructing 3D models was a painstaking and time-consuming task. These days, devices such as the Lynx A camera are turning it into a point and shoot affair. Microsoft is also doing its bit to make life easier for graphic artists, 3D printing aficionados, industrial designers, animators, architects, and games developers by turning its Kinect for Windows sensor to the task of generating 3D models of objects and environments.

Kinect Fusion, which was presented at SIGGRAPH in 2011 and was announced as a future update to the Kinect for Windows SDK in November last year, quickly creates highly detailed 3D models by taking depth data captured by the Kinect for Windows sensor and averaging the readings of sequences of hundred or thousands of frames. This allows the system to create a more detailed model than would be possible with a single reading.

3D models of large objects or environments can be generated by moving the Kinect sensor around said object or environment, while smaller objects can be scanned by moving the object around in front of the sensor. In addition to 3D model construction, Kinect Fusion can also be used for 3D augmented reality applications and taking 3D measurements.

"The amazing thing about this solution is how you can take an off-the-shelf Kinect for Windows sensor and create 3-D models rapidly," said Shahram Izadi, senior researcher at Microsoft Research Cambridge. "Normally when you think of Kinect, you think of a static sensor in a living room. But with Kinect Fusion, we allow the user to hold the camera, explore their space, and rapidly scan the world around them."

The technology, which was demonstrated this week at Microsoft’s annual TechFest technology fair, was developed by the Kinect for Windows team working in collaboration with Microsoft Research Cambridge.

Kinect Fusion can be seen in action in the video below.

View gallery - 4 images