基于Laws模板滤波的车标定位算法

Vehide-Logo Location Algorithm Based on Laws' Mask Filtering

针对车标受天气变化的影响而导致无法准确定位的问题,本文提出了基于Laws模板滤波的车标定位方法,根据车牌与车标的先验知识进行车标粗定位,再根据车标周围散热网的纹理信息,并利用Laws模板滤波除去车标周围散热网的干扰,实现车标的精确定位。试验结果表明,本文方法与Sobel边缘检测方法相比,可更好的保留车标纹理信息。该方法提高了车标定位的光照鲁棒性,定位结果更准确,而且对于较暗的车标也有很好的定位效果。

In this paper we come up with an algorithm of vehicle-logo location based on Laws＇ mask filte- ring in order to come over the question that vehicle-logo can＇t be localized accurately because of the light change. First, we locate the vehicle-logo coarsely according to the prior knowledge of the position relation- ship between the vehicle logo and the license plate;Then, according to the type of the attered network, we use the Laws＇s mask to filter out the background of the vehicle-logo and locate the accurate logo. The re- sult shows： compared with the Sobel edge detection , the method of this paper remains more texture infor- mation, and improves the algorithm robustness for the night, and at the same time, it has a good location effect for the dark vehicle logo.