Discuss the use of 'Optical Flow' in motion analysis.
Instruction: Explain what optical flow is and its application in analyzing object or camera motion.
Updated
Example Answer
Optical flow estimates apparent motion between images, usually as a two-dimensional displacement field. It can come from object motion, camera motion or both. Classical methods often assume approximate brightness consistency and locally smooth motion, so lighting changes, occlusions and large displacements can cause errors. I would use flow as one input to tracking or motion analysis. Converting it into metric velocity or collision risk needs camera geometry, timing, depth and additional uncertainty checks.
Make it your own
Illustrative failure case: a stationary wall produces substantial flow when the camera moves. Check camera motion and forward/backward consistency before interpreting that flow as object motion.
Why this works
This separates image displacement from physical motion. Follow-up: what information is missing when a flow vector is reported in pixels per frame?
Sources: OpenCV optical flow.