Re: Losing odometry when doing fixed point scanning

Posted by scanboy on
URL: http://official-rtab-map-forum.206.s1.nabble.com/Losing-odometry-when-doing-fixed-point-scanning-tp4336p4416.html

Thank you this was fairly easy to get going when you confirmed my thinking but slightly later in the code flow when doing Rtabmap::process(...) for the multi camera SensorData I get slightly confused.

The only overloaded functions that accept SensorData also requires a pose, and naturally I assume it would be for the entire multicamera object and at least when there is only one it wouldn't matter. I can see in older code (from around 2014) that process existed with only SensorData but not today. Digging past the v0.15 rtabmap.h file I see that just process(SensorData(image, id), Transform()) is used
But using only Transform() I get [ERROR] Rtabmap.cpp:945::process() RGB-D SLAM mode is enabled, memory is incremental but no odometry is provided. Image 0 is ignored!
Initializing the transform to something sensible like 0,0,1,0,  -1,0,0,0,  0,-1,0,0 gives cloud with <1000 points often < 400. (The data base viewer doesn't show anything valuable for this)

What could be a sensible init of the pose for process(...) for this multicamera object?

I'm loosely following the rgbddataset example.