Tailgating detection (also called piggybacking detection) flags when more than one person passes through a controlled entry on a single authorisation — a second person slipping through a door or turnstile right behind someone who badged in. It pairs video analytics with access control: the access system records one valid credential, the camera counts two bodies passing, and the mismatch becomes an alert. It targets one of the most common physical-security gaps, where the door is secure but people defeat it socially.
In practice it works by watching the choke point — the door, mantrap, or turnstile — and counting distinct people through it in a tight time window, then correlating that count with the access event. A turnstile or speed-gate enforces it physically; an open door relies on the analytic plus a human response. Integrated into the VMS and access platform, a tailgating event can pop the relevant camera onto the wall instantly so an operator can verify and respond.
The pitfalls are the geometry and the human factor. Close-following people, groups, carried objects, and shallow camera angles make accurate counting hard, so accuracy is a range that depends heavily on a good overhead or tight-angle view; busy entrances generate more ambiguity and more false alarms. And detection alone changes nothing without a response process — a tailgating alert that no one acts on is just a log entry. Design the camera placement for the count, integrate it with access control, and pair it with a real intervention plan.
