Application of artificial neural networks in number plate recognition system
Santrauka
We investigate the recognition of air number plates by Artificial Neural Networks aiming to make the system cheaper by the use of web cameras at the same time not sacrificing the performance and recognition rate. Main attention is devoted to the investigation of Kohonen Feature Map and Multilayer Perceptron, their structure and training parameters. The description of the system under construction, the composition of the recognition software module, simulation experiments, achieved recognition results and suggestions for the future development are given.