Happy new year , blabla ;p
Yesterday I took a video recording with my old camera that has an infrared “night-vision” mode..
The pattern projected seems to be static , and not very complicated , just a series of bright dots that change intensity , spacing and angle as you approach , or distance your self from the device due to physics..
The whole infrared “projector” is just a lens , so the cost of this part of the kinect couldn`t be over say 5 euros
I have 7 days until I return to the army , so I`ll better make them count..!
With the camera distance I have selected ( 6 cm away ) you have enough disparity to be able to perform good calculations , compared to the outside world video ( last post ) , where there is almost no disparity on far points of the view..
Nice thing to find out that you have shared goals with a gigantic organization such as Google xD
Its a good solution with zero processing cost , but Guarddog doesn`t need a Kinect to extract 3d Depthmaps ;P