Let’s train the program on random objects, like humans! What could possibly go wrong?
For crazy ideas wackos come up with.
We know where it isn’t (Xierumeng)
Sometimes, we don’t detect anything. What if we save all of those points and draw it on the map/graph as the places we know the tent isn’t? Then we can go fly somewhere else to look.
Seconded, I was thinking of this a few days ago. It would basically be an addition on the video mediator to a module that builds a map layer based on detected tent locations at that point.