Improved Algorithm for License Plate Location Based on Edge Detection and Mathematical Morphology

License plate localization is the crucial first step in the license plate recognition system. In this paper, the weighted average method was used to preprocess the license plate image, and the Roberts operator was used to preserve the contour information of the license plate area and the whole image.Combining mathematical morphology operations with the features of license plate images, using corrosion operations for filtering to achieve rough license plate localization.The precise location of the license plate area was realized by using the reasonable selection of structural elements in the closed operation and the open operation.This algorithm reduces the appearance of interference connected regions, thereby improving the speed of license plate localization.It effectively improves the traditional license plate location method and saves time for the subsequent license plate recognition.The experimental results show that this algorithm can accurately and quickly locate license plate images in different environments, and has certain robustness.