Localización automática de placas de automóviles

License plate localization for vehicle has been widely studied before, giving favorable results. However, as far as this research has advanced, there is no indication of extensive benchmarks to test the algorithms. Thus, in this work is presented the first stage of a vehicle’s license plate Automatic Recognition System, which is focused to only find the license plate. Three similar algorithms were developed using techniques as erosion, dilation, and Sobel filter, being tested in a benchmark with images from different illumination 127 Research in Computing Science 105 (2015) pp. 127–137; rec. 2015-09-06; acc. 2015-10-17 conditions, ambient problems, image size in pixels, among others. The results obtained shows that the best proposed algorithm allows successfully locate more license plates, supporting different image conditions, in comparison with the other algorithms, were one of them cannot detect the license plate in the majority of the images tested.