BriefCam License Plate Recognition White Paper

Introduction

License plate recognition (LPR) is a way to extract a vehicle’s license plate number from a video to a textual, searchable string. This allows the identification of vehicles of interest. LPR can be used in all of BriefCam’s modules: REVIEW, RESPOND and RESEARCH.

The accuracy of license plate recognition is dependent on many factors including camera location (placement), camera field of view, distance of camera from vehicles, resolution (pixels), exposure time, video quality, lighting, and angle of camera.

License Plate Recognition Flow

In BriefCam, license plate recognition is carried out as illustrated in the diagram below.

BriefCam first uses its vehicle detection abilities. In this way, BriefCam can focus its search for license plates only on the detected vehicles and not on the entire frame, resulting in more efficient processing.
A state-of-the art Deep Neural Network is used to detect the license place from the vehicle, dealing with issues such as edge detection and color transitions between the license plate and the car body.
BriefCam tracks the vehicle throughout its journey in the scene. The best frames for LPR transcription are then selected from the track.
The text is transcribed from the license plate.

This information can then be compared to a watchlist, searched manually, used as a reference when license plates are searched, or used for statistical data aggregation in RESEARCH.

License Plate Recognition Alerts

You can trigger rule-based alerts using license plate recognition watchlists.

If a vehicle in the video matches several entries in a watchlist or watchlists, a single alert will be triggered.

License plate recognition alerts do not have a cool-down period like face recognition alerts.

Disabling License Plate Recognition

The administrator can disable the License Plate Recognition feature in BriefCam. For more information, see Disabling FeaturesDisabling Features.

BriefCam White Papers