Automatic system for identification of a vehicle based on the application of the matrix QR code

Abstract

The article discusses a new method of automatic vehicle identification system (AVIS) based on the developed system of vehicle traffic control using matrix QR-code, which solves the problem of segmentation of license plates and signs of different formats (signs of the QR-code type with a different number of characters). The system has been created in order to recognize segmented characters from the same input sets that have the same size without overlapping letters and sets of numbers. It is proposed to apply a method for recognizing non-standard characters, based on the use of a mixed pattern method, the degree of compliance. Based on testing of 100 samples of templates and signs, experiments and images taken in real conditions of vehicle operation, it can be stated, that this method provides a recognition accuracy of 80.4%., for the mixed pattern method, it takes 1.7 seconds, and for the compliance templates it takes 0.75 second to perform the recognition operation when determining the adequacy of the application of the mixed method.