| With the rapid development of China economy, the number of cars has increaseddramatically.Highways and other infrastructure also are developing quickly, but can notmeet the increasing demand of cars.The huge number of vehicle management hasbecome a big problem in many countries.Using intelligent transportation system toimprove the management of automobile has been a tendency.Vehicle license plateautomatic recognition (ALPR) is an important part of the intelligent transportationsystem. Vehicle license plate recognition, based on the intelligent transportation systemthat with vehicle registration,authentication and monitoring alarm function, is widelyused in the parking lot,junctions,the charging system and so on.License plate recognition system generally consists of four important steps, that isimage pre-processing of the license plate, license plate location, character segmentationand character recognition. The license plate image pre-processing affects a direct impacton the final recognition result, and histogram equalization is an important part of thelicense plate image pre-processing. Therefore, the effect of histogram equalization willhave a final impact on the result of license plate recognition. While the traditionalhistogram equalization method will cause the gray-scale merger, the balanced imagewill produce false contours, then affecting the accuracy of license plate recognition.In this paper,to overcome the shortcomings of the traditional histogramequalization method, I put forward an improved histogram equalization method.Thisnew method which is better than traditional method in balance effect can retain all ofthe grayscale image after equalization,to effectively avoid the false contour. Newmethod can also adjust the weights to a different equilibrium effects, and differentgray-scale can be intensived. The new method is applied to license plate recognitionprocess, and ultimately achieved more ideal recognition results. |
PDF Full Text Request