Development of Vehicle License Plate Number Recognition Algorithm in Afghanistan

The increasing number of vehicles in all cities, not only makes it easy, on the contrary, it causes many problems to identify cars in various fields such as traffic control, management of parking lots, management of security areas, identification of unlicensed cars, etc There are cars whose licenses have expired, stolen cars, and cars with fake license plates in Afghanistan. Overcoming these problems requires the use of various financial and human resources, and human resources will not be able to manage and control the cars efficiently. By using an automatic system to recognize the car's license plate number, the problems can be overcome to a great extent. Although plate recognition algorithms and systems have been built in almost all advanced and developed countries, no system or any type of algorithm that is useful for Afghan vehicle plates has been developed yet. Vehicle plate number identification includes three basic steps, which include identifying the location of the plate; the extraction of plate characters; and the recognition of plate characters using the method of observation and library of all three mentioned steps. The activities performed in those areas have been investigated using the Matlab R2019B programming tool. Among the techniques used in the proposed algorithm are image monochromatic techniques, morphological methods to remove noise and edge detection, the use of histograms to increase the accuracy of plate location detection, the use of perceptual artificial intelligence with predetermined criteria for more accurate detection of the plate location, labeling to isolate the plate characters, and comparison with the original samples to identify the pointed plate characters. The results obtained from the research and implementation of our proposed algorithm are very promising and efficient, so the accuracy of detecting the location of the plate is 98% and the character recognition is 90%. Our method is better than any other proposed method or algorithm that was conducted during research.