Hello!
I'm currently trying to implement a camera-IMU calibration method in ATOM which requires a "denser" dataset than usual - meaning more collections - to identify points where the system transitions between stillness and motion using IMU data.
To record such a dataset, I thought of using the automatic_data_collector. However, this collector has hardcoded thresholds which it uses to identify when it should save collections to the dataset. If these thresholds were given as an argument, keeping the old hardcoded values as default, I think I could achieve what I need.
I also think this is going to be useful in further camera-IMU calibration implementations.
Tagging for visibility @miguelriemoliveira @manuelgitgomes!
Hello!
I'm currently trying to implement a camera-IMU calibration method in ATOM which requires a "denser" dataset than usual - meaning more collections - to identify points where the system transitions between stillness and motion using IMU data.
To record such a dataset, I thought of using the
automatic_data_collector. However, this collector has hardcoded thresholds which it uses to identify when it should save collections to the dataset. If these thresholds were given as an argument, keeping the old hardcoded values as default, I think I could achieve what I need.I also think this is going to be useful in further camera-IMU calibration implementations.
Tagging for visibility @miguelriemoliveira @manuelgitgomes!