Noise pattern of event cameras

[tonyde...@gmail.com]

Hi,

I am user of the DAVIS240. 

The noise issue is one of the problems to process event data. The noise pattern of a single event camera is determined by plenty of hardware parameters and environment changes. Would it is possible that one can provide me some paper to take an overview of which reason will produce these noises? As far as I have known, the temperature is indeed one of the reasons to lead to the noisy output. I was wondering if the lightness changes or motion conditions also lead to the change of noise pattern?

If I am wrong about anything, please correct me without hesitation.

Thank you and I am forward to hearing from you.

Tony

[Luca Longinotti]

Chapters 2 and 3 of:
https://arxiv.org/abs/2006.07722
go into detail on the noise sources and pixel operation.

For a detailed study of temperature effects, see:
https://ieeexplore.ieee.org/document/7962235

Different ligthing conditions, especially the amount of light, do
affect the pixel response (it is slower and noisier in darker scenes).
Motion should have less of a 'noise effect', but its amount will
influence the readout efficiency, possibly leading to some other non-
ideal behaviors like the striped readout pattern in DAVIS sensors at
high event rates. The DVXplorer and other more modern sensors are
generally less/not affected by having redesigned readout circuits.

Hope this helps, have a great week!
--
Luca Longinotti (llongi)

Software Engineer
iniVation AG - https://inivation.com/
Zurich, Switzerland
Office: +41 44 500 3214

Support - https://inivation.com/support/

[tonyde...@gmail.com]

Thanks, I appreciate your timely response.

Have a wonderful day!

Tony

[Tobi Delbruck (UZH-ETH)]

Thanks Luca, good pointer to the survey paper. Our V2E paper goes
further into noise: https://arxiv.org/abs/2006.07722

Tobi