Syntactic pattern recognizer for vehicle identification numbers

摘要

Road pricing is certain to be introduced in most cities by the year 2000. Most systems require vehicles to carry a transducer which identifies it. If this fails, the only way of identifying the vehicle is by reading the vehicle identification numbers (VIN). A recognition system is described which thins the characters using conventional algorithms to produce 8-connected skeletons and uses a new pattern grammar to describe these characters in terms of simple string. The same string represents a character irrespective of its position, size or orientation. The characters used to test the performance of the algorithm were extracted from a wide variety of images of vehicles. They are not idealized representations. The origin of these characters means that they are very varied in their size, orientation and completeness. No normalisation of characters is required prior to determining their representation in the pattern grammar described.