Optical flow is a way to understand how things move by looking at pictures. Imagine you are looking out of a car window and you see trees, buildings, people and cars moving by. Optical flow would try to understand how each thing in your view is moving relative to everything else.
Optical flow looks at two pictures taken at different times and tries to find the similarities and differences in these pictures. It finds where the same things are in both pictures and how much they have moved. In other words, it tries to understand how everything in the view has changed from one picture to the next.
It does this by looking at patterns of pixels in the two pictures. Each pixel is like a tiny dot of color in the picture. Optical flow looks at how these dots have changed from one picture to the next. It can then create a picture of arrows showing the direction and speed of how everything has moved.
We use optical flow in lots of things like checking for errors in web cameras or stabilizing drone cameras. It helps us understand how everything moves and how to make things look better.