Okay, your idea is feasible. You need to have Ffmpeg compiled for android. First have a look for it at Stack link.Then decide yourselves as per need.
After you have Ffmpeg compiled for android, you can just search and extract/add audio
as per your needs. To give a a start have a look at this and FFmpeg Docs Guide/Official Example.
To make video from images
have a look at Ffmpeg official example. You can find plenty of these on google.
After having all these things on your hand, you are ready for your project. I would suggest to try and familiarize yourself for Ffmpeg
on Windows/Linux as per your need first.
Hope this would help.
Cheers.:)