Deep Learning-Based Methods for Automatic License Plate Recognition: A Survey

The increasing need of a car to travel for our chores or as an essential mode of travel backup has become a kind of exigency. The large integration of information and technologies, underneath totally different aspects of the world, has led to the identification of cars as abstract resources in systems. With this, the need for a robust model that monitors the management systems, crime prevention has emerged. This can be often achieved by us or by special intelligent equipment which will enable the recognition of a car by its number plates in real environments. Various deep learning algorithms have proved to be useful for this cause. In this survey, we observe different methods of pre-processing, segmentation, detection, and classification for vehicle number plate detection using various machine learning and deep learning algorithms. We studied the different pre-trained models, with the layers and filters that help us efficiently identify the number plates in cars. This structured survey paper gives assistance for researchers who want to study and implement in this domain to understand the perspectives as well as issues. Future scope of this research paper can help us in designing an efficient algorithm with maximum accuracy as previous accuracies range up to 89–96%. With detailed analysis and application, we aim to increase precision.