Lists

Tuesday, June 09, 2020

3D News: Brookman, LG-Hitachi, Quanergy, Airy3D

Brookman publishes a video demonstrating its iTOF short-pulse BT008D sensor performance in 100Klux sunlight:




Hitachi-LG demos its 3D people tracker said to be installed in many stores:



Quanergy unveils its QORTEX People counting LiDAR:

"With over 98% accuracy, it will profile each individual without invading privacy and will detect tailgaters or someone It will detect tail gators or anyone coming by.

This particular platform is built with 100% OPA based solid state lidar.
"



Airy3D demoes its 3D sensor short range depth map capabilities:




13 comments:

  1. I am now Product Manager at Airy3D. Please comment here with any feedback or questions.

    ReplyDelete
  2. Do you have some impact on the color quality reproduction?

    ReplyDelete
    Replies
    1. No, there is only very minimal impact on color demosaicing, MTF, and QE with the Airy3D solution when properly optimized. You can basically consider the TDM (transmissive dispersive mask) as an extension of the microlens array and the CFA on the sensor.

      Delete
    2. Oops. The previous comment should say transmissive diffraction mask (TDM). I just saw that now. My apologies for the error.

      Delete
  3. There is no substantial effect on color/CFA, MTF, QE or overall image quality for the AIRY3D transmissive diffractive mask. Our patented technology and design know-how leverage the Talbot effect to optimize the sensor performance for low cost and ease of implementation with low computational burden.

    ReplyDelete
  4. What is the video frame rate if you implement your approach on modern smartphone 20MP+ resolution image sensor? Is it able to run on a smartphone processor or you need an ASIC or PC to process video?

    ReplyDelete
  5. Hi @Vladimir. Yes, there is indeed significant interest in designing the AIRY3D TDM technology into front cameras of smartphones. The AIRY3D solution benefits not just from the simple one-sensor design, but from neglibible effects on image quality and the low compute requirements suitable for low-power embedded processors.

    Depending on the specifics of the intended use case, there is a lot of flexibility in terms of the percentage of pixels used for generating the 3D point cloud data as well as the frame rate desired.

    As for other applications, these include gesture recognition, access control/smart home/IoT, and various industrial applications.

    ReplyDelete
    Replies
    1. This reply does not answer any part of the question :)

      Do you need an ASIC or can you run it on the smartphone AP?
      Any figures on expected FPS at given resolution?

      Please don't get me wrong, I believe this stuff is super exciting! For this reason we would love to see some figures to understand how it relates to ToF, Structred Light, Stereo and other techniques.

      Delete
    2. Contact me at first.last@airy3d.com and we can discuss further. There are a wide array of mobile and lightweight embedded processor options being explored, all of which have conventional frame rate expectations attached. Your suggestions on what is appropriate to your target applications would be most welcome and informative.

      Delete
  6. Can you share a raw output image of the sensor to get a feeling of what the TDM does?

    ReplyDelete
    Replies
    1. Check the article on AnandTech to get an idea, and check back for updates, since we also have some improvements to image quality to show off!

      Delete
    2. Here is the link to the AnandTech article I mentioned earlier. There is a side-by-side pair of images with and without the TDM. Furthermore, we are seeking to update the article to show more current capabilities.

      https://www.anandtech.com/show/15828/airy3ds-transmission-diffraction-mask-a-cheap-camera-depth-sensing-solution

      Delete
  7. New coverage on the AIRY3D solution today on Novus Light:
    https://www.novuslight.com/airy3d-introduces-depthiq-computer-vision-platform_N10604.html

    ReplyDelete

All comments are moderated to avoid spam and personal attacks.