Keywords: Camera networks, 3D reconstruction, RGB-D data, Multi Cloud Registration

This repository was built during my Msc Thesis, and it contains part of the code develop that it make possible to register a set of 3D camera, by using arucos.

In this repository there are two main functionalities:

1. Run python PosePipelineMaker.py Pipelines/cangalho_realsense.json in the passed file, there is information relative to the topic to capture the capturing mode and the ids of the present arucos.
2. Type an uppercase R to first calculate the rotations
3. Type an uppercase T to calculate the translations
4. The poses generated will be save in a file.

1. Run python PosePipelineMaker.py Pipelines/realsense_regular3 in the passed file, there is information relative to the topics to capture the capturing mode, the calibration object model and its ids.
2. Type an uppercase R to first calculate the rotations
3. Type an uppercase T to calculate the translations
4. The poses generated will be save in a file.

On top of these, the present other functionalities:

Run tfbroadcasterv2.py POSES_FILE where poses file is the file containing the camera poses retrieved from before. Opening rviz it is possible to see the camera poses in 3D space.

Run boxcaster.py and in another terminal rospccropper.py CAMERA_NAME, to crop a given pointcloud to only contain points within a box.

Run IntrinsicFetcher.py to obtain the intrinsic parameters of all the existing cameras