Motion detection refers to the process of detecting moving objects in a video or in consecutive image frames. It often involves background subtraction or frame differencing: if you have a static camera, you can estimate a background model and then any significant deviation in the current frame from that background is considered moving (foreground). Simple approaches might take pixel-wise differences between subsequent frames and threshold them to see where changes occur. More robust methods model each pixel’s intensity with a distribution (like Gaussian mixture models) to handle noise and gradual lighting changes. Motion detection can also involve optical flow (per pixel motion vectors) to understand motion dynamics. It’s commonly used in surveillance (detect intruders), video compression (detect moving areas to allocate bits differently), and as a first step in tracking or activity recognition systems. The output of motion detection is often a binary mask highlighting moving regions.