基于暗原色先验的雾天图像车牌定位

摘要

在雾天条件下，会影响到基于传统车牌识别方法的车牌识别系统的车牌识别准确率。首先根据He提出的暗原色先验原理应用IMF算法对图像去雾处理；然后进行小波边缘检测，定位车牌候选区域；最后根据车牌先验知识定位车牌区域。实验表明，该方法对雾天条件下车牌定位有良好的效果。%Fog weather conditions affect the license plate recognition system accuracy,which based on traditional license plate recognition method.Firstly,this paper employs Dark Channel Prior principle proposed by He to eliminate the influence of fog weather.Secondly uses b-spline wavelet edge detection to position license plate candidate area of restored images,final-ly locates the license plate according to the the candidate area.