Can anybody explain me why we use multiple videos while creating the project?

can anybody explain me why we use multiple videos while creating the project?
because even if i used 3 different videos when labeling the frames i was given the frames from only one video so iam little confused here.

You should label frames from different videos that span the space you wish to track in - different individuals, lighting, cameras, etc, so label each video in the GUI. Once you hit save an quit, it asks you if you want it label another folder.

Can we merge labeled dataset from different videos?

yes; when you create a training set, it merges the files.