ZEDGrids

Framework for Voxelgrid-based Space Monitoring in Human-Robot (UR5) Collaboration Environments, using ZED camera, OpenPose and PyntCloud. It allows to create several types of voxelgrids like the typical Occupancy Grid (binary) or our proposed Labeled Occupancy Grid, LOG, that labels voxels according to the elements present: human operator, robot, task objects, or static obstacles (like walls or tables).

Dependencies

Python Module:

• Python 3.7
• PyntCloud

C++ Module:

• ZED SDK
• OpenPose
• PCL
• OpenCV

Install:

Clone this repository :

    git clone https://github.com/DIGI2-FEUP/ZEDGrids

• Download and Install OpenPose
• Put the C++ Module files inside the OpenPose folder examples/user_code/
• Build OpenPose using CMake
• Install PyntCloud
• Place voxelgrid.py file from the Python Module (replace) in PyntCloud folder structures

Usage:

• Run voxel_pc_server.py

    python voxel_pc_server.py

• THEN Execute zed_openpose_pc.cpp (in Visual Studio)