The focal length is now fetched from the camera info topic. An f_mult_factor is also exposed through dyn.reconf. for easy adjustment of unwanted artifacts due to calibration etc.
We do need to have the max_depth_per_tile configurable by dynamic reconfigure because it is strictly an encoding coefficient responsible for including/discarding depth for various ranges. With a long range camera (~2m and above) this coefficient enables those ranges to be also encoded, otherwise the resulting depth will be lost. For more info regarding this parameter see #8 and https://github.com/RobotWebTools/ros3djs/pull/207
It also replaces the hardcoded 640x480 resolution to now support other resolutions (confirmed working well!) ;)
Adds a couple of parameters that can be tweaked during runtime and adds a possibility to fetch the focal length from a camera info topic.
From my description in the "internal" version we (will) use at Pick-it (https://github.com/Intermodalics/depthcloud_encoder/pull/1):
It also replaces the hardcoded 640x480 resolution to now support other resolutions (confirmed working well!) ;)