| License Plate Recognition(LPR) is one of the most critical research topics of Computer Vision, Image Processing and Pattern Recognition applying in Intelligence Transportation System. which is composed of Image Collection, Plate Location, Character Segment and Character Recognition. Plate Location is an important component of License Plate Recognition. thecolor,size and position of the plate in Plate Location is uncertain,and The interference of environmental factors, so Plate Location is a difficult problem. This paper studies a new method Based on Gray Level Jumping. Raise a concept of transition point density. First locate the license plate roughly, and then exactly locate the license plate.First, in order to eliminate thenoise,improve the quality of image and make the area of License Plate intensified.Image pretreatment includes image gray,histogram equalization and median filtering. Second, wo can find the trip points by first difference operator image, then decrease the trip points by threshold, estimate distance between the distance adjacent trip points and the density of trip points. And roughly locate the license plate based on the characteristic of license plate. Then, classify the license plate based on HSI color space. Especially,with an improved Ostu algorithm, binarization of the images is able to significantly reduce the effects of non-uniform illumination. Analyse 3 inclination'types of license plate. Align the license plate based on Hough transform and a new method of bilinear interpolation. CharacterSegmentis realized in the methods of pixel-value change and horizontal-vertical projection techniques combining with the inherence characteristics of the license area.The system of plate Location is constructed on the MATLAB platform which is required to verity the performance. The results prove that the accuracy of roughly locate the license plate could reach 93.08% and the accuracy of exactly locate the license plate could reach 90.80%. |
PDF Full Text Request