Samsung combines depth – and rgb-pixels on the image sensor

Samsung has a cmos image sensor is developed that is both in-depth information as rgb images. The resulting depth image has a resolution of 480 by 360 pixels, while rgb video with up to 1920 by 720 pixels is recorded.

The combination of a so-called time of flight depth sensor with normal rgb photodiodes makes it possible to have fewer components to use in electronics that both functions are needed. Such equipment may be cheaper. However, the new sensor as a limitation that the depth information and rgb-images are not recorded at the same time, but only sequentially, so we read with Tech-On.

Currently, the Kinect accessory for the Xbox 360 from Microsoft, the use of two dieptesensors and a single rgb camera movements in 3d to register. The dieptesensors in the Kinect, however, not use the time of flight method, but of the so-called structure light technology.

The pixels of the new image sensor from Samsung are built up from the normal rgb system, while for the in-depth information in a z-pixel across the full width of two rgb pixels is placed. This provides a resolution for the depth image of 480 by 360 pixels, while the rgb-part video with an unusual resolution of 1920 by 720 pixels.

The rgb subpixels are 2.5 µm broad and high, while the z-pixels with a size of 2.5 by 9µm. The color information that is missing by the integration of z-pixels, is completed by interpolation. To increase the efficiency of the z-pixels increase, Samsung has ‘stops’ placed between the light-sensitive surfaces of the rgb subpixels and the z-pixels.

The sensor is manufactured with a 0.13 µm process, and makes use of traditional beeldsensortechnologie with the metal bedradingslagen above the fotodiodes. Samsung thinks the quantumefficiëntie in a next phase to be able to double through the use of a bsi-cmos-process. Also can the rgb subpixels even better shielded from the z-pixels.