3D Scanning with Kinect v2

Finding an HD image alignment for the Kinect V2

By on Mar 24, 2015 |

depthgrid

Because the depth camera and color camera are different devices, the images they produce do not automatically overlap. The color camera will have different lensing effects than the depth camera. The further away from the center, the more pronounced the effect will be. Because each Kinect V2 has its own camera parameters (they are tested for and set at the factory), a simple fixed solution is not feasible–what will work perfectly on one specific Kinect may not yield the same results on another Kinect. Microsoft has included a feature to auto-align the images, however, this only provides a standard definition image–the one above has not been resized, it’s the full size output from the HD that has been ‘aligned’ with depth.

A little coding magic can grab additional pixel information and double the resolution as seen here (click to open in new window):

depthgrid_hires

Now to add some additional code to grab more data, to get rid of those grid lines.

depthgrid-touchup

And there you have it: an HD quality image that uses Microsoft’s alignment method. (I’ll clean up those left over white specs later.)

After a good nights rest, I cleaned up the process and here is the result:


hi-res_imagealignedtodepth

Note: Scan from Life is in no way affiliated with Microsoft. We are an entirely separate company that has created a product that is dependent on a Microsoft owned product.