Pattern Configuration for Brazilian License Plates

[Paulo A. L. Rego]

Dear Team,

I am currently using your ultimateALPR engine, and my work is focused on recognizing Brazilian license plates, which have two primary formats: the older standard (ABC-1234, pattern LLLNNNN) and the newer Mercosur standard (ABC1D23, pattern LLLNLNN).

I have read the documentation and understand that I should use the "charset": "latin" configuration, as this model is trained on a wide variety of plates, including those from Brazil.

My question is: to further improve recognition accuracy and minimize errors, is there a specific configuration parameter to define or prioritize a specific license plate pattern? For example, is there a key where I could specify that the engine should validate its results against the LLLNNNN and LLLNLNN formats?

Any clarification you could provide would be greatly appreciated.

Thanks

--

Paulo Rego