In modern systems of augmented reality revealed a fundamental problem. She is in a weak occlusion – that is, in a small blocking of light from distant objects to those that are located in the foreground. In common models of headsets, like the Microsoft HoloLens or Magic Leap, it takes so pronounced value and that creates serious problems.
If the object is in front of the eyes closes the light from the background, it looks more crisp and detailed – and Vice versa. Alas, the developers of augmented reality originally came from the fact that the drawn objects must have some degree of transparency. Modern headsets use only one spatial light modulator, which is not enough. Therefore, the added objects because of low occlusion at times so transparent that they are barely visible – about any realistic perception of the prospects in this case speech does not go. Accordingly, to interact with the object, which really can not see, is also very problematic.
Simple solution, adding the second light modulator, immediately changes the whole hardware architecture and significantly complicates and increases the device headset. Therefore, at Stanford University has gone a different way – a team of engineers under the leadership of brook Kranich developed a system of mirrors that act as a filter for light from the real objects. These mirrors have two positions: full pass light and full locks, and each tiny mirror can be switched independently of the others.
The main advantage of development – the shift speed is measured in tens of thousands of times per second. The controller selects that position of the mirrors to ensure the best scheme for the darkening picture, to block the light of real-world objects in favor drawn. This task requires a lot of calculations because the system will have to add powerful processor, plus will increase power consumption. But it will be able to keep the overall dimensions of set in their current format.
Source — IEEE