Page 43/89
VIDAR
User Manual
Adaptive Recognition
America
Adaptive Recognition
Hungary
Adaptive Recognition
Nordic
Adaptive Recognition
Singapore
SENDINFO@ADAPTIVERECOGNITION.COM
WWW.ADAPTIVERECOGNITION.COM
Typically, 3-8 images are first received by a Preselection stage of the pipeline, where a very fast
system selects the best image for plate recognition. A preliminary set of license plate data is already
available after this stage. The image best representing the event is then forwarded to the ANPR
stage, where a slower, but much more thorough analysis determines the final and complete set of
license plate data (license plate text, position, nationality, etc.). The user can configure multiple
ANPR stages in series. An example: the first ANPR stage is equipped with an ANPR engine that
recognizes only domestic plate types (in a country or a state), thus it is very fast (e.g. a IRL/GBP
engine). This stage will recognize the majority of the vehicle plates; the corresponding events will
pass through the next stage. This next, second ANPR stage is equipped with a broader scope engine
(e.g. a latin engine), which is less not quite as fast, but will recognize the remaining (not domestic)
plates. Also an MMR stage can be inserted in the pipeline, which adds make and model info to the
event. The result is then stored in a database. The database is accessed through a web server, with
structured queries in 'Pull' mode. This is demonstrated in the Browse menu. Alternatively, in 'Push'
mode, the camera can forward data using a given protocol (HTTP, FTP, SFTP). This is configured at
the Result Upload page.
Summary of Contents for Vidar
Page 1: ......