Funding entity Ministerio de Industria Turismo y Comercio
Participated by INDRA SISTEMAS S.A., SICUBO S.L., ITINERE INFRAESTRUCTURAS S.A.
The launch of this project arises from the need to develop an automated system capable of extracting features of the cars from images taken on the road, in order to improve the accuracy of automatic recognition as much as possible. Despite the widespread use of automatic number plate recognition, there are a number of limitations in several elements that make up such systems, which yield to small but undesirable level of recognition errors.
The proposed platform integrates new image acquisition devices, semantic analysis of features, clustering and classification based on neural networks, and a database management system. The main objective is to improve the capabilities of automatic recognition to the point of obtaining results in situations in which human recognition is not able to read the plate correctly. When plate is not completely available in the image of the car, the system would be able to identify the vehicle if there is a collection of images from the same vehicle available in the system database, reasonable situation for vehicles that use the infrastructure frequently.
The project involves a preliminary analysis, functional development and implementation of a prototype that demonstrates the degree of reliability attained by the platform.