Thursday, October 12, 2017

Intel Unveils D400 Realsense Camera Family

Intel publishes an official page of D400 camera family, currently consisting of D415 and D435 active stereo cameras. Reportedly, the earlier Realsense cameras SR300, R200 and F200 are being discontinued, while D400 series will be expanded to include passive and active stereo models:


  1. How come they do not report any depth sensing resolution with such detailed specifications tables? Are they really providing 3D depth map cameras?

  2. Did any of Intel's previous Realsense solutions make it into products or do they keep releasing better and faster products year over year without actual sales?

    1. In this video from a year ago, Intel Realsense VP says that they've sold "millions" of 3D cameras:

      They are used in a couple of laptops that implement Windows Hello feature - a sort of Apple Face ID, just few years earlier:

      This is one example:

    2. No blockbusters, but here a few
      Yuneec Typhoon H (
      Amazon echo look (
      Asus zenbo (
      Razer Stargazer (

  3. Can anybody explain how exactly active IR stereo differs from structured light?

    1. Regular stereo cameras need texture or high contract details in the scene to extract the depth. They can't measure distance to a plain white wall, for example. The active illumination adds its pattern to create texture for the places where there is no natural one.

      Structured light system has just one camera that looks on the projected light pattern only while ignoring the scene natural texture.

    2. Thanks Vladimir.
      So we can say that active IR stereo is also projecting some IR pattern but it does process natural scene texture in addition to the IR pattern.

      In this case, isn't active IR stereo in general more advantageous than structured light that ignores the scene texture? I would assumen that with IR active stereo, you can get better depth data in outdoor applications.

    3. Active stereo and structured light cameras treat the projection pattern in a different way. Stereo cam looks at disparities between the two cameras. Structured light camera looks for the distortions and shifts in the projected pattern resulting from the distance changes.

      As for the advantages, it depends on your criteria. It one values power consumption and price above all, the structured light like one in iPhone X is better. On the other hand, stereo works better in bright sunlight, especially at large distance.

      Of course, both stereo and structured light systems have zillions of variations that improve on their weak points and make the whole picture quite complicated.

    4. Theoretically, structured light is active stereo. Two cameras - stereo, camera and projector - active stereo. Disparity computed the same in both cases. Two cameras and projector would be 3 view active stereo.
      The phase shifting is one of the most common used methods that is not based on disparity computation.
      It is a bit sad that Intel has only implemented Gray code structured light in the MEMS based RealSense, projector hardware should be capable for phase shift method. It's harder computationally, but produce much higher accuracy and needed less number of frames per scan.

  4. they promote their products via imaging companies like Framos and try to sell it more also in industrial environment:


All comments are moderated to avoid spam and personal attacks.