Not sure how to explain this use case but tried my best.
We are recording the molten flow through jetson nano - CSI camera.
we need to identify the shape of the flow and compare with ideal flow to ensure that there is no leak or overflow. The challenge is that flow shape is not consistent due to launder which keep moving up/down in which the flow happens and hence different shapes. How do we identify the shape of moving object and compare with ideal image?
I know the solution is more related to the technique we should practise in opencv but posting here to seek help.