Optical flow is the capability of a moving autonomous vehicle or robot to perceive and predict the motion of nearby objects by analyzing their apparent movement within its camera’s field of view. When the robot moves, stationary objects appear to shift in the opposite direction at a speed related to its own motion. Conversely, moving objects display different trajectories and speeds.
By interpreting these relative motions, an autonomous robot can estimate the direction and velocity of other moving objects, while also factoring in depth information and its own movement. This insight is crucial for safe navigation, collision avoidance, and dynamic path planning in real time. Optical flow relies entirely on visual data captured by cameras and sophisticated algorithms in computer vision.