Estudio comparativo entre algoritmos de reconocimiento de borde para identificación de placas de autos.

Abstract

Comparative study of Edge Recognition Algorithms for identification license plates in the Department of Systems and Telematics in ESPOCH. This research is supported in the method "Deductive" that allowed the same statistical analysis based on a procedure that is to develop a theory and then deducing its consequences and the method "Comparative" has permitted a systematic comparison between objects of study by determining the most efficient algorithm to identify the characters found on a license plate. Sources of information was taken from web pages, comments, photographs, and materials used are Hardware: Computer, IP Camera and Software: DBMS MySQL, QT Creator, OpenCV to determine the guidelines that should be covered. The results obtained based on the parameters are compared processing time: Sobel algorithm has a 37 millisecond response against Canny algorithm with 40 ms; successes: Sobel algorithm presents a percentage of 81% and 85% Canny algorithm. It can be concluded that implementation of the Canny algorithm is more efficient compared with Sobel algorithm, its efficiency is not just reflected in the response time, but on the quality of edge detection. We recommend the use of the Sobel algorithm for efficiency in response time and quality of edge detection for image processing libraries of "OpenCV" to be made freely available significantly reduces implementation costs, without forgetting excellent use of hardware resources.