Análisis de rendimiento de algoritmos de reconocimiento de placas de números de vehículos desarrollado mediante la transformación de ondas discretas y la correlación de imagen digital

Abstract

Objective: To develop two algorithms to evaluate through the comparison the performance of the Discrete Wavelet Transform (DWT) and Digital Image Correlation (DIC) techniques for the recognition of patterns, and more specifically the automatic recognition of license plates. Results y Conclusions: It can be concluded that one algorithm outperforms the other depending on the test performed; it is imperative to conclude that the most optimal algorithm for the recognition of vehicle registration using the DWT in conjunction with the DIC is that in which the recognition with the HOG detector is performed as an FBM technique; because although the algorithm that uses the normalized cross-correlation has the best performance in terms of time processing.

Keywords

• reconocimiento automático de matrículas
• transformada wavelet discreta
• correlación de imágenes digitales
• histograma de gradientes orientados
• correlación cruzada normalizada