Use Case to support:
Capturing of objects like machines or cars with a 360° exterior video (walkround) using an iPhone.
The user will capture the object with a smartphone, not using any stabilizer device / gimbal. Therefore the algorithm has to 'normalize' the sequence of pictures in such a way that the frames are centralized relative to each other. A light photogrammetry model could be the solution. As it is a 360° walkaround the algorithm could make a reference from frame to frame, but also adjusting the frames corresponding to each other, i.e. the front side and backside of a viewing angle (example: a car view from the drivers side corresponds with the view from the passengers side or in other words the frames of 0° and 180° could correspond ... at least in a car use case)
The sequence of pictures as a 360° walkaround shall be produced in real time, i.e. the performance of the model is an important criteria.
A sample 360° walkaround can be provided to illustrate the tasks.
PLEASE PROVIDE EXAMPLES OF SIMILAR TASKS SUCCESSFULLY EXECUTED