针对雨天、雾霾天等复杂天气条件,提出了一种车牌定位方法.首先采用CLAHE方法进行图像增强并控制噪声放大,利用维纳滤波去除非周期噪声.通过对雨天图像频域特征进行分析,设计出一种新的利用频域掩模滤除雨水条纹的方法.最后结合边缘检测与形态学运算,提取车牌区域,并利用Radon变换进行倾斜校正.实验证明,文中提出的车牌定位算法适于复杂天气情况下车辆图像的定位,定位准确率达到96%,具有一定的实用价值.%In view of the complicated weather conditions such as rainy days,fog and haze,a method of license plate location is proposed.Firstly,the CLAHE method is used to enhance the image and to control the noise amplification,the wiener filtering is used to remove aperiodic noise.Based on the analysis of the frequency domain characteristics of rain image,a new method of filtering rain streaks using frequency domain mask is designed.Finally,combined with edge detection and morphological operations,the license plate region is extracted and the Radon transform is used for tilt correction.Experimental results prove that the license plate location algorithm proposed in this paper is suitable for vehicle image localization in complex weather conditions,and the localization accuracy is 96%.It has certain practical value.
展开▼