Steps
Phase 1: Test different ways of obtaining the 3D data.
sensors (Kinect V1 / V2)
software techniques : Fusion libraries for manual coding which can be done in C# or C++, testing out 3rd party software and comparing results. Single shot + stitching together vs continuous scanning.
rotating the person vs rotating the camera around a static person
Once these experiments have been concluded we can move onto the next phase
Phase 2: Skeletal data export
capturing a snapshot of the skeletal data for the subject and exporting this to a meaningful format
determine the distance that the Kinect sensor needs to be in order to get the best skeletal reading
determine a file format that makes sense for the remaining steps of the project
These are somewhat straightforward objectives with clear outcomes so its just a question of obtaining the data and exporting successfully.
Phase 3: Camera rig and person platform control
a camera rig that supports the kinect and allows for the device to move so that we can capture all the required data with ease
platform that can be controlled by a motor to provide smooth and continuous movement
positioning of each Kinect to ensure that we capture all the required data both depth and skeletal
At this stage we should have a physical artefact that allows us to capture the data that we need successfully.
Phase 4: Putting it together
This is one of the hardest parts, mapping the skeletal data to the











