Pure cpp example? about kiss-icp

[nachovizzo]

Hello @JaySlamer, yes, this is a problem indeed. Now we have a mixture between C++/Python. But from next week on, I will start working in a proper ROS wrapper that does not enforce python as dependency. Will this work in your case? Or do you think it is still valuable to have a C++ CLI?

I will keep this issue open as a reminder to push for the ROS wrapper

PS: The "big" missing part on the C++ is only the adaptive threshold estimation, but if you use a fixed threshold(that also works pretty well) then all the ingredients are already on the C++ side.

[huawei-sai]

@nachovizzo Thank you for this excellent work.

Having a ROS wrapper with only C++ implementation would make easy to integrate with existing pose-graph back ends (https://github.com/gisbi-kim/SC-A-LOAM/blob/main/src/laserPosegraphOptimization.cpp). I would vote for this and will be waiting to test it myself.

The main reason to use ROS wrapper is to run the SLAM systems on-device easily as most of us use ROS for communicating the data throughout the robot, both front-end and other perception modules.

[nachovizzo]

@JaySlamer Not yet there, but here is a draft on how to use the system only with C++

[JaySlamer]

Hello @JaySlamer, yes, this is a problem indeed. Now we have a mixture between C++/Python. But from next week on, I will start working in a proper ROS wrapper that does not enforce python as dependency. Will this work in your case? Or do you think it is still valuable to have a C++ CLI?

I will keep this issue open as a reminder to push for the ROS wrapper

PS: The "big" missing part on the C++ is only the adaptive threshold estimation, but if you use a fixed threshold(that also works pretty well) then all the ingredients are already on the C++ side.

A ROS wrapper would be great. Thank you.

[chengwei0427]

It's a good idea for a ROS wrapper with only C++ implementation.

I'll check and update.

Thank you.