Object tracking
Also known as: subject tracking, visual tracking, motion tracking
Following a chosen object or person through video frames, so software knows where it is at each moment. Used by drones, gimbals, editors and AI tools.
Updated
What it means
Object tracking starts with a target, either selected by the user or found by a detector, and then locates it in every following frame. It must handle changes in size, angle, lighting and partial blocking, such as a rider passing behind a tree or turning away from the camera.
Drones use tracking to follow subjects in modes like ActiveTrack, gimbals use it to keep a face centered, and editing software uses it to attach text or blur to moving objects. AI editing tools use tracking to understand who the main subject is across a clip.
Tracking relies on computer vision, often combined with motion prediction to guess where the object will appear next.
How it works in practice
For live tracking on drones and gimbals, help the system with contrast: wear clothing that stands out from the background, avoid crowds of similar subjects and keep the subject reasonably large in frame. Spotlight style modes also rely on this.
In editing, motion tracking lets you attach a label, a graphic or a blur to a moving rider or license plate. Start the track on a clear frame and correct drifting points manually.
Tracking quality drops with motion blur and low light, so fast shutter speeds and good light help both live and post tracking.
What to watch out for
Trackers lose the subject during occlusion, when something passes in front, and may jump to a similar-looking object when the subject reappears. Always watch tracked results, and take over manually when needed, rather than trusting the system blindly.
Fast, erratic motion such as crashes, spins and quick camera pans can exceed what the tracker can follow, especially at lower frame rates and in poor light.
Live tracking on drones involves safety risks if the system fails near obstacles. Keep the drone in visual line of sight and ready for manual control.
How RawClip handles it
RawClip uses AI analysis to understand motion, faces and peak action in your footage, then cuts the best moments to music. Upload clips from every camera, and the AI builds one highlight.