The Research Of License Plate Location Algorithm Based On Mathematical Morphology And Granularity Computing Of Quotient Space Theory

License plate recognition system (LPR) can through the ingestion vehicles image, to the image which gathers carry on processing and the analysis automatically, and can distinguish the car license according to the car license apriori knowledge and other information. The car license localization's main purpose is in the car license image determined that the vehicles license plate's concrete position, and will contain the vehicles license plate character the sub-image to divide from the entire vehicles license plate, for the following character division and recognition utilization.Quotient space theory, simulation of different size according to people the ability to think, to discuss on the fields, properties, structure, said under the different particle size, nature and those that the interdependence between nature, the relationship between conversion, the problem on a variety of different size space (commercial space) are analyzed, and then draw on the comprehensive solution of the original problem, improve problem solving efficiency.Therefore, this article proposed that one kind feasible based on the morphology and business space granularity computation theory's solution mentality is:Carries on the division pretreatment first using mathematics morphology theory to the image, the union textural property analysis, the use business space granularity theory of dissociation, does not intersect the target background region car license image dissection each independence; Then using business space's granularity synthesis principle, carries on the synthesis to each business space, thus obtains the car license region exact location.The people in the long-term social practice process, knew that the actual problem formed gained the information method from certain different angles, passed through certain logical thinking again, utilized the rich imagination, will obtain the information integration obtained the complete concept. Based on business space granularity synthesis's taxonomic approach is precisely based on this kind of thought. This article main idea including the following content:1 st, we uses the morphology knowledge, carries on the pretreatment to the car license. The car license image's binaryzation good trades the degree immediate influence car license localization the effect, this article based on the auto-adapted initial threshold value binaryzation algorithm, proposed one kind of improvement's auto-adapted initial threshold value binaryzation algorithm, analyzed through the experiment, obtained the good processing effect.2nd, to already obtained the picture element region to correspond to the different textural property information car license region pictorial information classification, these category message has then formed from the different angle to the car license region image new observation, corresponds the different granularity space these different new observations, the use business space granularity synthesis principle, the merge different texture classification result question may transform as to the different granularity space synthesis question. And guarantees each business space which according to business space granularity synthesis theory the false principle will obtain to organize, observes obtains to the image new understanding, realizes the car license region final classification, then realizes the car license region final localization finally.3rd, and guarantees the false principle based on business space granularity computation's granularity synthesis principle, the majority of backgrounds are pared, the car license doubtful region already indicated that in the image was only left over the minority connective quite good region, the connection carriage analytic method has applied in the car license two value images, and used 8- connection mark method and 8- connection euler number computation, the judgment car license real position.