best FPS for motionless scene?

[EmmanuelCos]

| Camera Model | D435 |

Hello,
I am using a D435 cam to capture motionless scenes, and I am wondering what would be the best FPS to use in this case? Are there any other parameters that are recommended to use when capture motionlessscene, to improve depth quality and precision?

Thank you and regards,

Emmanuel

[agrunnet]

In general the exposure time should be dictated by the lighting. So in normal lighting this may be about 5ms for the d435. That means that you can actually set the frame rate to 90fps and there is no benefit to going to shorter frame rate. However with a motionless scene you should definitely use the post-processing temporal filter and you can decrease the alpha parameter to increase the filtering. This will improve the results about 3x.

[MartyG-RealSense]

For optimal depth accuracy the 848x480 resolution should be used for D435.

If you find that there is noise or holes in the image then you can improve the quality of the image by applying post-processing filters to perform enhancement functions such as hole-filling, as @agrunnet mentioned above.

If possible, avoid fluorescent lights such as ceiling strip lights, which may flicker at frequencies that are hard to see with the human eye and create noise on the image.

[EmmanuelCos]

Thank you for your answers! And the exposure time for the depth cameras shouldn't be increased when there is no motion?

[EmmanuelCos]

I am trying to play with the temporal filter on realsense viewer, modifying the alpha parameter, but don't see any difference at all. Should I reset the cameras or so?

[MartyG-RealSense]

@agrunnet can give better advice on exposure times than I can. I would add the following points though:

• An exposure of 5 ms for normal lighting is suggested above. In Intel's excellent camera tuning guide for the 400 Series cameras, which @agrunnet co-authored, Section 15 on the bottom of the final page states that in bright sunlight the exposure can be reduced to near 1 ms (which can also reduce motion artifacts).

https://www.intel.com/content/dam/support/us/en/documents/emerging-technologies/intel-realsense-technology/BKMs_Tuning_RealSense_D4xx_Cam.pdf

• Manually setting exposure may have an effect on FPS.

#1957 (comment)

In regard to making changes in the Viewer to the temporal filter, I did run tests with different values and found that the image did not appear to change much if changing the temporal alpha value whilst the camera was motionless. The difference was most apparent when moving the camera, with the image taking longer to adjust and re-stabilise after a camera movement if the temporal alpha value was set low, whilst the image updated faster after a camera motion if a higher value was used.

Below, the upper picture illustrates slow update of the camera image after a motion with a small temporal alpha value, whilst the lower picture shows a less jagged image due to faster update because of a high alpha value.

[MartyG-RealSense]

Hi @EmmanuelCos Do you still require assistance with this case please, or can it be closed? Thanks!

[MartyG-RealSense]

Case closed due to no further comments received.