Video Analytics for Photobooths: Lost Items, Cleanliness and Telegram
Busted works with a photobooth network present in more than 100 countries. We detect left-behind items and dirt inside booths and send Telegram alerts. This geographic reach describes the client's network. Below are project frames and an example system message.
What needs attention inside a photobooth
A visit may leave a bag, printed photo strip or dirt on the floor. The service team needs to know where to check and see the context. We address this through image analysis and Telegram notifications.
The supplied frames highlight a bag on the seat, a photo strip and a dirty floor section. A box indicates the area flagged by the system. Staff can assess the image and decide what action is needed.
What the alert looks like
The Telegram message contains an image, frame time and a description of the change. In the supplied example, the system compares the interior with a reference and flags possible damage near a tiled-floor edge.
This screenshot comes from an RTSP stream test. The message also notes that a curtain obscures part of the floor and full scene inspection is not confirmed. This tells staff which area was available for analysis and whether further inspection is needed.
Validating a similar scenario
Deployment defines the observation zone, types of lost items and dirt, acceptable scene changes and recipients. Validation should include a clean booth, ordinary visits, lighting changes and a curtain blocking the view.
Useful metrics include detected and missed events, unnecessary and duplicate messages, alert latency and staff review time. Evaluate these on annotated episodes under the actual location's conditions.
Starting a similar project
Prepare before-and-after frames, service rules and camera details. Agree which events need a response and who receives messages. The initial review can establish pilot scope and validation data.
