The Research And Implementation Of Vehicle License Plate Location Technology Based On Adaboost

License plate location is one of the key technologies of automatic license platerecognition. The problem of poor positioning slow and positioning accuracy for license platelocation and license plate color are characterized by the license plate location method basedon Adaboost. This essay has a very important engineering application.This essay first analyzed the domestic and international research situation anddevelopment trend of the license plate location. The speed of positioning for the license plateis slow, due to the background light interference. This essay raised the real-time location ofthe Adaboost license plate based on color features principles and methods. On the basis of thecolor characteristics of the license plate area, the CCD color image sensor took images. Thelicense plate (called the target) in the analysis of plate area image and license plate near thearea (referred to as background) was analyzed. The image of license plate was convertedinto color model. That means RGB image was converted into the HSV image. With thefeatures of license plate images of characters and the license plate background having a fixedcolor, the license plate image binarization operation has been done. The license plate area washighlighted. And then the mathematical morphology processing algorithms was applied toremove noise in the binary image, to exclude most of the counterfeit license plate target. Therough location of the license plate was done. Then, the procedures of the license plateautomatic identification of the target image was prepared, based on the Adaboost algorithm.Calibration software was prepared to aim for an image for the blue background and whitecharacters. The license plate of the original image was collected. Based on the useful life, aswell as light intensity and other factors, image was filtered, calibrated and normalized. Thepositive and negative sample library of car license plate was established. Then the sampleoffline training was made by using Haartrianing in the OpenCV source library. The40×12license plate classification was made. Base on these, an accurate location of license plate wasmade after the rough location which was done by license plate classification. The geometryspace coordinates of the license plate was abtained. Finally, experiments were done. Theresults showed that the method has many features such as faster positioning, positioning accuracy and anti-background light interference.