I have been bias tuning the kria embedded kit with the imx636 sensor and have reached a limit of around 1.8-1.9m for pose updates. I have also adjusted focus to ensure I can pick out the LED's at that distance and checked in metavision viewer for both bias settings and focus. I can clearly see the LED blinking pattern and have almost eliminated all background/scene noise.

Close to this sub-2 meter distance the tracks required begin to diminish (min of 4 for planar marker) and therefore the pose estimator is unable to reliably update pose.

I understand that LED intensity diminishes with the square of distance, but I just want to confirm that with the current prophesee supplied LED marker board + imx636 (tuned bias and adjusted focus) the max theoretical distance to compute pose is no more than 2 meters.

If this is the case, the only remedy is to create a custom marker board with larger LED's at higher intensity (and a larger marker board). Using this research paper as a guide, I think they use a 9cm length marker board and 8x 1 Watt LED's to get 10m range. So the marker detection algorithm and IMX636 is not the limitation according to that paper. I am using a different lens than the paper, but not too far off from their spec lens.