Open SasinPhimsiri opened 5 months ago
Hi @SasinPhimsiri,
Those cameras have overlaps, so I took one reference camera (which will be scale 1), and I picked a vehicle in the beginning of the video and one at the end, and checked the elapsed time in seconds. Now if we do the same on other cameras, and:
I calculated the offsets similarly, by picking the appearance of the same vehicle in the beginning of the video (say $t_1$ on the reference. and at $t_2$ on the other camera), then the offset is $t_1 - t_2$.
If the cameras don't have overlaps, it's more difficult I guess. Probably one can do it by picking a few vehicles and checking the differences in their transition times. (which is needed anyway for configuring the transition windows)
Can you describe step by step on cityflow_s01 and cityflow_s02? Thank you :)