Introduction

What does the Vehicle Detector Agent do?

The Vehicle Detector Agent preselects the best images for ANPR (number plate recognition) and MMR (make, model, and color recognition) of vehicles from a video stream.

Who is it for?

Users of the Adaptive Recognition Vehicle API should use the Vehicle Detector Agent to prevent images that are not relevant for ANPR (images containing no vehicles or multiple frames of the same vehicle) from unnecessarily taking up processing power and bandwidth.

How does it work?

The agent detects and tracks vehicles in the input video stream and automatically selects the best image frame for each vehicle. It then transfers this single image to the API, captures the response and

  • forwards it to an HTTP POST webhook (if configured),
  • stores and serves it via a REST API (only the last 5000 vehicles are stored) and
  • displays the input stream and the list of recognized vehicles on a web dashboard.