Can't imagine that would be terribly difficult. They don't move like large insects for sure.
You'd think in that case one large insect would move a lot faster than the other.
If it can track it continuously, I imagine speed will pretty much clear up any confusion.
Yeah, continuously is the point here. The frequency of the data points is largely dependent of the distance to the object. And you need to be able to identify the object in two subsequent images.
I guess there's not that many insects at high altitudes either.