Blurry phenomenon in some areas of the event frame

[李翘楚]

Recently, I have been conducting research on stripe projection imaging technology using event cameras. The surfaces of the objects being photographed are matte.
Now, I have encountered a problem. The pixels in a certain area near the center of the event frames obtained through event stream synthesis are all very blurry. I don't know if it is a hardware issue of the event camera itself or some other problem. When collecting and processing event frames under different lighting conditions and angles, this blurriness issue persists.
I have sorted out the event stream slices that constitute the corresponding event frames, with a slice interval of 1 ms.

[Luca Longinotti]

Hello, it looks like the blurriness is the result of there being less events in that region (more sparse output).

Is it always the same region of the image that's affected? Looks mostly in the center?

These kind of issues (locally consistent artifacts) often have two major causes:

1) the light itself: for some reason that region has less light, or more flickering, or light is reflected differently resulting in less photons going towards the camera. Maybe more/stronger light, or increasing the sensitivity of the sensor, can help.

2) the lens: the type and quality of the lens can cause distortions and artifacts, and especially for certain applications, the interplay with the sensor's micro-lenses and their CRA (chief ray angle) can create areas where light is more concentrated or scattered.
For more details see: https://commonlands.com/blogs/technical/lens-chief-ray-angle-and-mismatch

Our DVXplorer line of cameras, based on the Samsung VGA sensor, has a micro-lens array with a CRA optimized for smartphone lenses (see under 'non-linear Mobile' in the link above), so if used with normal lenses in certain specific applications these kind of effects can be more pronounced. I've attached the CRA profile of the S5K231Y sensor.

For the DAVIS346, all micro-lenses are centered on their pixel, basically giving a uniform 0° CRA over the whole array.

Hope this helps, have a nice week!