對 intel realsense F200 和 Kinect V1的codingpattern原理進行調研，知二者都使用lightcoding的方式，該方法與傳統的coded structure light不同：傳統編碼結構光（空間編碼，時間編碼）最終使用的是三角測量的原理計算深度資訊；lightcoding通過投射具有高度偽隨機性的鐳射散斑，會隨著不同距離變換不同的圖案，對三維空間直接標記，通過觀察物體表面的散斑圖案就可以判斷其深度。

1. （Components）Optical apparatus, including first and second diffractive optical elements(DOEs

2. （Size）The first DOE applies, such as an uncorrelated pattern to input beam; the second DOE project pattern onto surface over a wide angular range. In Kinect V1, it creates 15 beams, in a 3x5 fan-out, the second DOE also can be 3x3, 5x5, 7x7 or other fan-out patterns.

3. （Properties）The phase mask

4. （Power）The shape of the phase mask within each cell of second DOE controls the power distribution among the different output beams. It can be designed to give either uniform light intensity distribution among tiles or a predefined non-uniform intensity distribution.

下面給出Kinect V1在不同距離產生的鐳射散斑圖案：

0.5m

1.0m

1.5m

2.0m

2.5m