Sunday, April 14, 2013

Review of 3D Imaging Approaches published a popular review of 3D imaging techniques: stereo, many versions of structure light and ToF. The authors list is impressive: Michael Brading (Aptina), Kenneth Salsman (Aptina), Manjunath Somayaji (Aptina), Tim Droz (Softkinetic), Daniƫl Van Nieuwenhove (Softkinetic), Pedro Gelabert (TI), and Brian Dipert (Embedded Vision Alliance and BDTI).

Naturally, the review ends with a comparison table:


  1. Finally the best solution should be stereo+structured light.

  2. The outdoor performance of time of flight also depends on the specifics of the hardware. It appears that most commercial implementations use differential electronics to perform background subtraction, but even so, bright sunlight can saturate pixel sensors before the subtraction occurs. Also I'd be interested if the authors are aware of the Softkenetic cameras which are much less than $1000 USD.

  3. We are aware of the fact that Kinect is sold at around 100$.

  4. now Softkenetic not surport outdoor
    pmd and panasonic surport outddor

  5. The Kinect is from Microsoft (designed by primesense? which isn't a ToF system like the SoftKinetic cameras.

    I've heard a talk by a leader in the field that thinks that the Kinect is a structured light system, but I have read a paper from primesense that describes the use of laser speckle decorrelation and pattern matching. What the Kinect actually does is unclear.

    1. Kinect uses an LED based projector with a random pattern. In a certain way - this is structured light. However, this technology is single shot and thus capable of motion imaging.

      Traditional structured light requires multiple patterns and images. This leads to lower frame rate and problems with moving objects.


All comments are moderated to avoid spam.