USE OF GENETIC ALGORITHMS TO AUXILIATE VEHICLE PLATE DETECTION IN IMAGE SEGMENTATION USING OPENCV

Resumo

One of the fundamental steps of image processing is the segmentation of images, which
aims to partition the image in a way that makes data analysis and extraction simple. In this article, a
method is proposed that performs the segmentation of images, using multiple thresholds, through
genetic algorithms to find the contours of a car's plate. For this, the algorithm should validate a set of
characteristics that define a region as being a car plate. Identify optimal threshold parameters in order
to obtain the best possible solution for image segmentation and detection of vehicular plaques. The
methodology used consists of the bibliographic review of an image segmentation article on Genetic
Algorithm Application for the Evolution of Parameters for Image Segmentation and an article on the
detection of vehicular plaques using Python and OpenCV. The employee image bank was obtained
from a public repository of images. The parameters that will be estimated to validate whether or not a
given object is detected on a board are: the board's outline and the presence of alphanumeric
characters. In this way, it will be possible to optimize detection efficiency by eliminating false positive
results, such as image contours that do not correspond to the plates. The segmentation of the image through multi-thresholds occurred and it was possible to identify vehicle plates in the image with
greater precision.