Abstract
License plate recognition (LPR) is a well studied problem first attempted in the 1970s. The image analysis part in LPR renders the problem hard to solve perfectly, leading to continued research. A defining step is the thresholding of the image to separate an object in the image from its background. While considerable effort has been devoted to LPR, thresholding is hardly ever mentioned, always assuming an effective one. A large number of thresholding methods exist. These methods perform differently in various image analysis problems but rather poorly if applied in an automatic fashion to the LPR problem. We introduce a new thresholding method that performs well in automatic LPR. The method uses colour-based histogram classification information. The approach leads to a 93.4% success rate and can be applied to the automatic recognition of symbols and characters.