The Research On Car License Location Which Has Adjustment Function For Inclines And Character Division

Along with the modernization of the city day by day, the quantity of the automobiles all over the world is increasing rapidly. This causes burden of the highway traffic to be increasingly heavy, and caused serious attention to the highway traffic automatic control and the management and the so-called intelligence transportation system (ITS Intelligence System) in many countries. And gradually applies in many domains such as the transportation information statistics collection, the road workshop correspondence, the parking lot management, no stopping automatic charge vehicles automatic travel and so on. But above domain are all related with the license plate automatic recognition. License plate recognition (LPR), is not only realized in the transportation street intersection, the highway, the military fort, the institution gate guards to the passing vehicles real-time registration, the current capacity statistics and to defends the goal the safe garrison, the vehicles which cause trouble, the robbed vehicles and the crime vehicles are correctly recognized and Intercepted. But also may register and inquire the vehicles in the automobile parking which passes in and out. So that safeguards safely prevents to steal, and May assist automatically record the expense, this has saved the man power, and the fund without doubt, simultaneously enhanced the traffic control efficiency at the same time.Along with the computation and the performance enhancement and the computer vision technology development, this kind of car license automatic diagnosis technology was already mature day by day, the system was consummated day by day. But there still has some problems when it is applied in actual center. For example, because the particularity when absorb the car license, which usually causes inclines to the car license. But the inclined license is very difficultly divided and recognized; therefore the solution to the inclined license is extremely essential.Generally speaking, the overall system divides into four major parts. Such as picture gathering, the car license localization, the character division, and the character recognition. This article researches and analyzes the existing theories and algorithms,Had expounded a series of pretreatment process in detail such as gradation, the binaryzation, the edge examination and the filter of the gathered image and so on under the Visual c++ platform; In the car license localization module, This article is in the foundation of using the edge examination and binaryzation method,Using the gradation jump characteristic, examines the up-down boundary of car license. Then in this foundation, use vertical projection method examines the left-right boundary of the car license. Then, the car license was located completely. In the module of rectifying of inclined car license, this inclined image is adjusted by the average height statistics of the left half and right half; Finally, the aritcal make use of rough divide to get a character's height firstly, and then subdivide to getting single character's width, in this foundation the system may realize the following development, the last step of character recognition. Under this premise had simultaneously pointed out the system insufficiency and waits for the enhancement direction.