Vehicle-License-Plate Image Denoising And Image Restoration Based On Multiscale Method

Intelligent traffic management system is the development trend of21st century worldroad traffic.Correct identification of vehicle not only achieve the statistics and query on trafficflow and measurement and management of road load,but also provide proof for accidentvehicle,smuggled vehicles,and missing vehicles.The light has a great impact on image qualityin vehicle image,and different lighting conditions and illumination angle also has a greateffect on unevenness of illumination of license plates.Different time,climatic conditions,background light,reflective degree of license plate determine the brightness features of thelicense plate.As the development of computer technology,processing problem of vehicle-license-plate image has got more and more attention and research.In this article,multiscale inversion algorithm is constructed with the combination ofmultigrid method,and restoration and denoising of blurred image is carried out,makingmultiscale inversion method identify the characteristics of the original image and greatlyreduce the numerical workload.First collected vehicle-license-plate images are transferredinto grayscale images.Because of the similar brightness of the resulted gray image of targetand background,the image gray stretch is necessary.While the traditional method is linearstretch and hard-threshold,multi-scale stretch of the image and soft-threshold based on fuzzymathematics is applied to restore the original image information without losing original imagefeatures.The results of numerical simulation show that the effectiveness of the multigrid onimage denoising and restoration under certain conditions.