(2023) Silva, Diego Saragoza da; Kabbani, Gabriel Mauricio; Loureiro, Luís Filipe Martins; Rodrigues, Theo Barbara
This project has the objective of developing an API that reads images or videos, identifies
license plates, blurs them and extracts their information. As output, the user receives the
characters of the license plate as well as the same image or video with the license plates blurred,
for anonymization purposes. To achieve identification and data extraction from the license
plates, it was necessary to study image processing techniques, machine learning approaches,
text extraction algorithms and information about APIs architecture. Results showed that the
usage of a machine learning model together with object tracking techniques led to an accuracy
of more than 86%. In terms of character extraction, the group achieved an accuracy of 57%.
